1 2017-01-26 Martin Sebor <msebor@redhat.com>
3 * gimple-ssa-sprintf.c (format_floating): Test HAVE_XFmode and
4 HAVE_DFmode before using XFmode or DFmode.
5 (parse_directive): Avoid using the z length modifier to avoid
6 the ISO C++98 does not support the ‘z’ gnu_printf length modifier.
9 * gimple-ssa-sprintf.c (adjust_for_width_or_precision): Change
10 to accept adjustment as an array.
11 (get_int_range): New function.
12 (struct directive): Make width and prec arrays.
13 (directive::set_width, directive::set_precision): Call get_int_range.
14 (format_integer, format_floating): Handle width and precision ranges.
15 (format_string, parse_directive): Same.
17 2017-01-26 Jakub Jelinek <jakub@redhat.com>
20 * dwarf2out.c (generate_skeleton_bottom_up): For children with
21 comdat_type_p set, just clone them, but keep the children in the
25 * dwarf2out.c (prune_unused_types): Mark all functions with DIEs
26 which have direct callers with -fvar-tracking-assignments enabled
28 (resolve_addr): Avoid adding skeleton DIEs for DW_AT_call_origin
31 2017-01-26 Martin Sebor <msebor@redhat.com>
34 * gimple-ssa-sprintf.c (struct result_range): Add likely and
36 (struct format_result): Replace number_chars, number_chars_min,
37 and number_chars_max with a single member of struct result_range.
39 (format_result::operator+=): Adjust.
40 (struct fmtresult): Remove bounded. Handle likely and unlikely
42 (fmtresult::adjust_for_width_or_precision): New function.
43 (fmtresult:type_max_digits): New function.
44 (bytes_remaining): Handle likely and unlikely counters.
45 (min_bytes_remaining): Remove.
46 (format_percent): Simplify.
47 (format_integer, format_floating): Set likely and unlikely counters.
48 (get_string_length, format_character, format_string): Same.
49 (format_plain, should_warn_p): New function.
50 (maybe_warn): Call should_warn_p. Update diagnostic messages
51 and handle those for all directives, including plain strings.
52 (format_directive): Handle likely and unlikely counters.
53 Remove unnecessary quoting from diagnostics. Add an informational
56 (pass_sprintf_length::compute_format_length): Simplify.
57 (try_substitute_return_value): Handle likely and unlikely counters.
59 2017-01-26 Carl Love <cel@us.ibm.com>
61 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Remove
62 bogus entries for the P8V_BUILTIN_VEC_VGBBD built-ins
64 2017-01-26 Vladimir Makarov <vmakarov@redhat.com>
67 * lra-assigns.c (setup_live_pseudos_and_spill_after_risky): Take
68 endianess for subregs into account.
69 * lra-constraints.c (lra_constraints): Do risky transformations
70 always on the first iteration.
71 * lra-lives.c (check_pseudos_live_through_calls): Add arg
72 last_call_used_reg_set.
73 (process_bb_lives): Define and use last_call_used_reg_set.
74 * lra.c (lra): Always continue after lra_constraints on the first
77 2017-01-26 Kirill Yukhin <kirill.yukhin@gmail.com>
79 * gcc.target/i386/avx512bw-kshiftlq-2.c: Use unsigned long long
81 * gcc.target/i386/avx512bw-kshiftrq-2.c: Ditto.
83 2017-01-26 Jakub Jelinek <jakub@redhat.com>
85 * config/i386/avx512fintrin.h (_ktest_mask16_u8,
86 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): Move to ...
87 * config/i386/avx512dqintrin.h (_ktest_mask16_u8,
88 _ktestz_mask16_u8, _ktestc_mask16_u8, _kadd_mask16): ... here.
89 * config/i386/i386-builtin.def (__builtin_ia32_ktestchi,
90 __builtin_ia32_ktestzhi, __builtin_ia32_kaddhi): Use
91 OPTION_MASK_ISA_AVX512DQ instead of OPTION_MASK_ISA_AVX512F.
92 * config/i386/sse.md (SWI1248_AVX512BWDQ2): New mode iterator.
93 (kadd<mode>, ktest<mode>): Use it instead of SWI1248_AVX512BWDQ.
95 2017-01-26 Marek Polacek <polacek@redhat.com>
98 * fold-const.c (operand_equal_p) [COND_EXPR]: Use OP_SAME_WITH_NULL
99 for the third operand.
101 2017-01-26 Jakub Jelinek <jakub@redhat.com>
104 * omp-low.c (struct omp_context): Add simt_stmt field.
105 (scan_omp_for): Return omp_context *.
106 (scan_omp_simd): Set simt_stmt on the non-_simt_ SIMD
107 context to the _simt_ SIMD stmt.
108 (lower_omp_for): For combined SIMD with sibling _simt_
109 SIMD, make sure to use the same decls in _looptemp_
110 clauses as in the sibling.
112 2017-01-26 David Sherwood <david.sherwood@arm.com>
115 * gimplify.c (omp_notice_variable): Add GOVD_SEEN flag to variables in
118 2017-01-26 Jakub Jelinek <jakub@redhat.com>
121 * reg-stack.c (emit_swap_insn): Instead of fld a; fld b; fxchg %st(1);
122 emit fld b; fld a; if possible.
124 * brig-builtins.def: Update copyright years.
125 * config/arm/arm_acle_builtins.def: Update copyright years.
127 2017-01-25 Michael Meissner <meissner@linux.vnet.ibm.com>
130 * config/rs6000/vsx.md (vsx_extract_<mode>_store): Use wY
131 constraint instead of o for the stxsd instruction.
133 2017-01-25 Carl Love <cel@us.ibm.com>
135 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Fix order
136 of entries for ALTIVEC_BUILTIN_VEC_PACKS and P8V_BUILTIN_VEC_VGBBD.
138 2017-01-25 Jonathan Wakely <jwakely@redhat.com>
140 * doc/invoke.texi (C++ Dialect Options): Fix typo.
142 2017-01-25 Richard Biener <rguenther@suse.de>
144 PR tree-optimization/69264
145 * target.def (vector_alignment_reachable): Improve documentation.
146 * doc/tm.texi: Regenerate.
147 * targhooks.c (default_builtin_vector_alignment_reachable): Simplify
149 * tree-vect-data-refs.c (vect_supportable_dr_alignment): Revert
150 earlier changes with respect to TYPE_USER_ALIGN.
151 (vector_alignment_reachable_p): Likewise. Improve dumping.
153 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
156 * config/arm/arm.md (xordi3): Force constant operand into a register
159 2016-01-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
161 * doc/invoke.texi (-fstore-merging): Correct default optimization
162 levels at which it is enabled.
163 (-O): Move -fstore-merging from list to...
166 2017-01-25 Richard Biener <rguenther@suse.de>
169 * omp-expand.c: Include debug.h.
170 (expand_omp_taskreg): Make sure to generate early debug before
171 outlining anything from a function.
172 (expand_omp_target): Likewise.
173 (grid_expand_target_grid_body): Likewise.
175 2017-01-25 Maxim Ostapenko <m.ostapenko@samsung.com>
178 * asan.c (get_translation_unit_decl): New function.
179 (asan_add_global): Extract modules file name from globals
180 TRANSLATION_UNIT_DECL name.
182 2017-01-24 Eric Botcazou <ebotcazou@adacore.com>
185 * config/arm/arm.c (arm_function_ok_for_sibcall): Add back restriction
186 for long calls with APCS frame and VFP.
188 2017-01-24 David Malcolm <dmalcolm@redhat.com>
190 * cfg.c (original_copy_tables_initialized_p): New function.
191 * cfg.h (original_copy_tables_initialized_p): New decl.
192 * cfgrtl.c (relink_block_chain): Guard the call to
193 free_original_copy_tables with a call to
194 original_copy_tables_initialized_p.
195 * cgraph.h (symtab_node::native_rtl_p): New decl.
196 * cgraphunit.c (symtab_node::native_rtl_p): New function.
197 (symtab_node::needed_p): Don't assert for early assembly output
199 (cgraph_node::finalize_function): Set "force_output" for __RTL
201 (cgraph_node::analyze): Bail out early for __RTL functions.
202 (analyze_functions): Update assertion to support __RTL functions.
203 (cgraph_node::expand): Bail out early for __RTL functions.
204 * final.c (rest_of_clean_state): Don't call delete_tree_ssa for
206 * function.h (struct function): Update comment for field
208 * gimple-expr.c: Include "tree-pass.h".
209 (gimple_has_body_p): Return false for __RTL functions.
210 * Makefile.in (OBJS): Add run-rtl-passes.o.
211 * pass_manager.h (gcc::pass_manager::get_rest_of_compilation): New
213 (gcc::pass_manager::get_clean_slate): New accessor.
214 * passes.c: Include "insn-addr.h".
215 (should_skip_pass_p): Add logging. Update logic for running
216 "expand" to be compatible with both __GIMPLE and __RTL. Guard
217 property-provider override so it is only done for gimple passes.
219 (skip_pass): New function.
220 (execute_one_pass): Call skip_pass when skipping passes.
221 * read-md.c (md_reader::read_char): Support filtering
222 the input to a subset of line numbers.
223 (md_reader::md_reader): Initialize fields
224 m_first_line and m_last_line.
225 (md_reader::read_file_fragment): New function.
226 * read-md.h (md_reader::read_file_fragment): New decl.
227 (md_reader::m_first_line): New field.
228 (md_reader::m_last_line): New field.
229 * read-rtl-function.c (function_reader::create_function): Only
230 create cfun if it doesn't already exist. Set PROP_rtl on cfun's
231 curr_properties. Set DECL_INITIAL to a dummy block.
232 (read_rtl_function_body_from_file_range): New function.
233 * read-rtl-function.h (read_rtl_function_body_from_file_range):
235 * run-rtl-passes.c: New file.
236 * run-rtl-passes.h: New file.
238 2017-01-24 Jeff Law <law@redhat.com>
240 * config/microblaze/microblaze.h (ASM_FORMAT_PRIVATE_NAME): Increase
243 2017-01-24 Bin Cheng <bin.cheng@arm.com>
245 PR tree-optimization/79159
246 * tree-ssa-loop-niter.c (get_cst_init_from_scev): New function.
247 (record_nonwrapping_iv): Improve boundary using above function if no
248 value range information.
250 2017-01-24 Pekka Jääskeläinen <pekka@parmance.com>
251 Martin Jambor <mjambor@suse.cz>
253 * brig-builtins.def: New file.
254 * builtins.def (DEF_HSAIL_BUILTIN): New macro.
255 (DEF_HSAIL_ATOMIC_BUILTIN): Likewise.
256 (DEF_HSAIL_SAT_BUILTIN): Likewise.
257 (DEF_HSAIL_INTR_BUILTIN): Likewise.
258 (DEF_HSAIL_CVT_ZEROI_SAT_BUILTIN): Likewise.
259 * builtin-types.def (BT_INT8): New.
260 (BT_INT16): Likewise.
261 (BT_UINT8): Likewise.
262 (BT_UINT16): Likewise.
263 (BT_FN_ULONG): Likewise.
264 (BT_FN_UINT_INT): Likewise.
265 (BT_FN_UINT_ULONG): Likewise.
266 (BT_FN_UINT_LONG): Likewise.
267 (BT_FN_UINT_PTR): Likewise.
268 (BT_FN_ULONG_PTR): Likewise.
269 (BT_FN_INT8_FLOAT): Likewise.
270 (BT_FN_INT16_FLOAT): Likewise.
271 (BT_FN_UINT32_FLOAT): Likewise.
272 (BT_FN_UINT16_FLOAT): Likewise.
273 (BT_FN_UINT8_FLOAT): Likewise.
274 (BT_FN_UINT64_FLOAT): Likewise.
275 (BT_FN_UINT16_UINT32): Likewise.
276 (BT_FN_UINT32_UINT16): Likewise.
277 (BT_FN_UINT16_UINT16_UINT16): Likewise.
278 (BT_FN_INT_PTR_INT): Likewise.
279 (BT_FN_UINT_PTR_UINT): Likewise.
280 (BT_FN_LONG_PTR_LONG): Likewise.
281 (BT_FN_ULONG_PTR_ULONG): Likewise.
282 (BT_FN_VOID_UINT64_UINT64): Likewise.
283 (BT_FN_UINT8_UINT8_UINT8): Likewise.
284 (BT_FN_INT8_INT8_INT8): Likewise.
285 (BT_FN_INT16_INT16_INT16): Likewise.
286 (BT_FN_INT_INT_INT): Likewise.
287 (BT_FN_UINT_FLOAT_UINT): Likewise.
288 (BT_FN_FLOAT_UINT_UINT): Likewise.
289 (BT_FN_ULONG_UINT_UINT): Likewise.
290 (BT_FN_ULONG_UINT_PTR): Likewise.
291 (BT_FN_ULONG_ULONG_ULONG): Likewise.
292 (BT_FN_UINT_UINT_UINT): Likewise.
293 (BT_FN_VOID_UINT_PTR): Likewise.
294 (BT_FN_UINT_UINT_PTR: Likewise.
295 (BT_FN_UINT32_UINT64_PTR): Likewise.
296 (BT_FN_INT_INT_UINT_UINT): Likewise.
297 (BT_FN_UINT_UINT_UINT_UINT): Likewise.
298 (BT_FN_UINT_UINT_UINT_PTR): Likewise.
299 (BT_FN_UINT_ULONG_ULONG_UINT): Likewise.
300 (BT_FN_ULONG_ULONG_ULONG_ULONG): Likewise.
301 (BT_FN_LONG_LONG_UINT_UINT): Likewise.
302 (BT_FN_ULONG_ULONG_UINT_UINT): Likewise.
303 (BT_FN_VOID_UINT32_UINT64_PTR): Likewise.
304 (BT_FN_VOID_UINT32_UINT32_PTR): Likewise.
305 (BT_FN_UINT_UINT_UINT_UINT_UINT): Likewise.
306 (BT_FN_UINT_FLOAT_FLOAT_FLOAT_FLOAT): Likewise.
307 (BT_FN_ULONG_ULONG_ULONG_UINT_UINT): Likewise.
308 * doc/frontends.texi: List BRIG FE.
309 * doc/install.texi (Testing): Add BRIG tesring requirements.
310 * doc/invoke.texi (Overall Options): Mention BRIG.
311 * doc/standards.texi (Standards): Doucment BRIG HSA version.
313 2017-01-24 Richard Biener <rguenther@suse.de>
316 * ipa-devirt.c (odr_types_equivalent_p): Fix typo in diagnostic.
318 2017-01-24 Martin Jambor <mjambor@suse.cz>
321 * ipa-prop.c (ipa_free_all_node_params): Call summary destructor.
322 * ipa-prop.c (ipa_node_params_t::insert): Initialize fields known_csts
325 2017-01-24 Aldy Hernandez <aldyh@redhat.com>
328 * gimple-ssa-warn-alloca.c (alloca_call_type): Make sure
329 casts from signed to unsigned really don't have a range.
331 2017-01-24 Markus Trippelsdorf <markus@trippelsdorf.de>
333 * gimple-ssa-sprintf.c (format_floating): Change MPFR_RNDx to
334 GMP_RNDx for compatiblity.
336 2017-01-24 Martin Liska <mliska@suse.cz>
339 * tree-ssa-reassoc.c (rewrite_expr_tree_parallel): Insert assert
340 that would prevent us to call alloca with -1 as argument.
342 2017-01-24 Jakub Jelinek <jakub@redhat.com>
344 * dwarf2out.c (output_compilation_unit_header, output_file_names):
345 Avoid -Wformat-security warning.
347 2017-01-23 Andrew Pinski <apinski@cavium.com>
349 * config/aarch64/aarch64.c (thunderx2t99_addrcost_table): Improve
352 2017-01-23 Martin Sebor <msebor@redhat.com>
355 * gimple-ssa-sprintf.c (warn_level): New global.
356 (format_integer): Use it here and throughout the rest of the file.
357 Use the same switch to compute sign as base.
358 (maybe_warn): New function.
359 (format_directive): Factor out warnings into maybe_warn.
360 Add debugging output. Use warn_level.
361 (add_bytes): Use warn_level.
362 (pass_sprintf_length::compute_format_length): Add debugging output.
363 (try_substitute_return_value): Same.
364 (pass_sprintf_length::handle_gimple_call): Set and use warn_level.
367 * gimple-ssa-sprintf.c (struct format_result): Remove constant member.
368 (struct fmtresult, format_integer, format_floating): Adjust.
369 (fmtresult::fmtresult): Set max correctly in two argument ctor.
370 (get_string_length, format_string,format_directive): Same.
371 (pass_sprintf_length::compute_format_length): Same.
372 (try_substitute_return_value): Simplify slightly.
375 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust formatting.
376 (fmtresult::operator+=): Outlined.
377 (struct fmtresult): Add ctors.
378 (struct conversion_spec): Rename...
379 (struct directive): ...to this. Add and remove data members.
380 (directive::set_width, directive::set_precision): New functions.
381 (format_percent): Use fmtresult ctor.
382 (get_width_and_precision): Remove.
383 (format_integer): Make naming changes. Avoid computing width and
385 (format_floating): Same. Adjust indentation.
386 (format_character, format_none): New functions.
387 (format_string): Moved character handling to format_character.
388 (format_directive): Remove arguments, change return type.
389 (parse_directive): New function.
390 (pass_sprintf_length::compute_format_length): Move directive
391 parsing to parse_directive.
393 2017-01-23 Jakub Jelinek <jakub@redhat.com>
395 * tree.h (assign_assembler_name_if_neeeded): Rename to ...
396 (assign_assembler_name_if_needed): ... this.
397 * tree.c (assign_assembler_name_if_neeeded): Rename to ...
398 (assign_assembler_name_if_needed): ... this.
399 (free_lang_data_in_cgraph): Adjust callers.
400 * cgraphunit.c (cgraph_node::analyze): Likewise.
401 * omp-expand.c (expand_omp_taskreg, expand_omp_target):
404 2017-01-23 Richard Biener <rguenther@suse.de>
406 PR tree-optimization/79088
407 PR tree-optimization/79188
408 * tree-ssa-threadupdate.c (mark_threaded_blocks): Move code
409 resetting loop bounds after last path deletion. Reset loop
410 bounds of the target loop, make code match the comments.
411 * tree-ssa-threadbackwards.c (pass_early_thread_jumps::execute):
412 Make sure loops need no fixups.
414 2017-01-23 Kelvin Nilsen <kelvin@gcc.gnu.org>
416 * config/rs6000/rs6000-builtin.def (VSIEDPF): Add scalar insert
417 exponent support with double type for first argument.
418 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Changed
419 type returned by __builtin_vec_extract_sig,
420 __builtin_vec_extract_sig_sp, and __builtin_vec_extract_sig_dp
421 functions from "vector int" to "vector unsigned int" or from
422 "vector long long int" to "vector unsigned long long int".
423 Changed type returned by __builtin_vec_extract_exp,
424 __builtin_vec_extract_exp_sp, and __builtin_vec_extract_exp_dp
425 functions from "vector int" to "vector unsigned int" or from
426 "vector long long int" to "vector unsigned long long int".
427 Changed return type of __builtin_vec_test_data_class,
428 __builtin_vec_test_data_class_sp, and
429 __builtin_vec_test_data_class_dp from "vector int" to
430 "vector bool int" or from "vector long long int" to "vector bool
431 long long int" and changed second argument type from "unsigned
432 int" to "int". Added new overloaded function forms "vector float
433 __builtin_vec_insert_exp (vector float, vector unsigned int)" and
434 "vector float __builtin_vec_insert_exp_sp (vector float, vector
435 unsigned int)" and "vector double __builtin_vec_insert_exp (vector
436 double, vector unsigned long long int)" and "vector double
437 __builtin_vec_insert_exp_dp (vector double, vector unsigned long
438 long int)". Changed return type of
439 __builtin_scalar_test_data_class and
440 __builtin_scalar_test_data_class_sp and
441 __builtin_scalar_test_data_class_dp from "unsigned int" to "bool
442 int" and changed second argument from "unsigned int" to "int".
443 Changed type returned by __builtin_scalar_test_neg,
444 __builtin_scalar_test_neg_sp, and __builtin_scalar_test_neg_dp
445 from "int" to "bool int". Added new overloaded function form
446 "double __builtin_scalar_insert_exp (double, unsigned long long int)".
447 * config/rs6000/vsx.md (xsiexpdpf): New insn for scalar insert
448 exponent double-precision with floating point first argument.
449 * doc/extend.texi (PowerPC AltiVec Built-in Functions): Adjust
450 documentation of scalar_test_data_class, scalar_test_neg,
451 scalar_extract_sig, scalar_extract_exp, scalar_insert_exp,
452 vector_extract_exp, vec_extract_sig, vec_insert_exp, and
453 vec_test_data_class built-in functions to reflect refinements in
454 their type signatures.
456 2017-01-23 Andreas Tobler <andreast@gcc.gnu.org>
458 * config/aarch64/aarch64.c (aarch64_elf_asm_constructor): Increase
460 (aarch64_elf_asm_destructor): Likewise.
462 2017-01-23 Bernd Schmidt <bschmidt@redhat.com>
464 PR rtl-optimization/78634
465 * config/i386/i386.c (ix86_max_noce_ifcvt_seq_cost): New function.
466 (TARGET_MAX_NOCE_IFCVT_SEQ_COST): Define.
467 * ifcvt.c (noce_try_cmove): Add missing cost check.
469 PR rtl-optimization/71724
470 * combine.c (if_then_else_cond): Look for situations where it is
471 beneficial to undo the work of one of the recursive calls.
473 2017-01-23 Bin Cheng <bin.cheng@arm.com>
475 PR tree-optimization/70754
476 * tree-predcom.c (stmt_combining_refs): New parameter INSERT_BEFORE.
477 (reassociate_to_the_same_stmt): New parameter INSERT_BEFORE. Insert
478 combined stmt before it if not NULL.
479 (combine_chains): Process refs reversely and compute dominance point
482 2017-01-23 Martin Liska <mliska@suse.cz>
484 PR tree-optimization/79196
485 * tree-ssa-strlen.c (fold_strstr_to_memcmp): Rename to ...
486 (fold_strstr_to_strncmp): ... this. Fold the pattern to strncmp
488 (strlen_optimize_stmt): Call the renamed function.
490 2017-01-23 Michael Matz <matz@suse.de>
492 PR tree-optimization/78384
493 * tree-ssa-loop-split.c (patch_loop_exit): Use correct edge.
495 2017-01-23 Richard Biener <rguenther@suse.de>
497 PR tree-optimization/79186
498 * tree-vrp.c (register_new_assert_for): Make sure we've seen
499 both incoming edges before moving an assert.
501 2017-01-23 Martin Jambor <mjambor@suse.cz>
503 * ipa-prop.c (load_from_param_1): Removed.
504 (load_from_unmodified_param): Bits from load_from_param_1 put back
506 (load_from_param): Removed.
507 (compute_complex_assign_jump_func): Removed stmt2 and just replaced it
508 with stmt. Reverted back to use of load_from_unmodified_param.
510 2017-01-23 Martin Jambor <mjambor@suse.cz>
513 * ipa-prop.h (ipa_param_descriptor): Anotate with with GTY(()).
514 (ipa_node_params): Annotate with GTY((for_user)). Make descriptors
515 field a pointer to garbage collected vector, mark lattices and
516 ipcp_orig_node with GTY((skip)).
517 (ipa_get_param_count): Adjust to descriptors being a pointer.
518 (ipa_get_param): Likewise.
519 (ipa_get_type): Likewise.
520 (ipa_get_param_move_cost): Likewise.
521 (ipa_set_param_used): Likewise.
522 (ipa_get_controlled_uses): Likewise.
523 (ipa_set_controlled_uses): Likewise.
524 (ipa_is_param_used): Likewise.
525 (ipa_node_params_t): Move into garbage collector. New methods insert
527 (ipa_node_params_sum): Annotate wth GTY(()).
528 (ipa_check_create_node_params): Adjust to ipa_node_params_sum being
530 (ipa_load_from_parm_agg): Adjust declaration.
531 * ipa-icf.c (param_used_p): Adjust to descriptors being a pointer.
532 * ipa-profile.c (ipa_profile): Likewise.
533 * ipa-prop.c (ipa_get_param_decl_index_1): Likewise.
534 (ipa_populate_param_decls): Make descriptors parameter garbage
536 (ipa_dump_param): Adjust to descriptors being a pointer.
537 (ipa_alloc_node_params): Likewise.
538 (ipa_initialize_node_params): Likewise.
539 (load_from_param_1): Make descriptors parameter garbage collected.
540 (load_from_unmodified_param): Likewise.
541 (load_from_param): Likewise.
542 (ipa_load_from_parm_agg): Likewise.
543 (ipa_node_params::~ipa_node_params): Removed.
544 (ipa_free_all_node_params): Remove call to delete operator.
545 (ipa_node_params_t::insert): New.
546 (ipa_node_params_t::remove): Likewise.
547 (ipa_node_params_t::duplicate): Adjust to descriptors being a pointer,
548 copy known_csts and known_contexts vectors.
549 (ipa_read_node_info): Adjust to descriptors being a pointer.
550 (ipcp_modif_dom_walker): Make m_descriptors field garbage
552 (ipcp_transform_function): Make descriptors variable garbage
555 2017-01-23 Andrew Senkevich <andrew.senkevich@intel.com>
557 * config/i386/avx512bwintrin.h: Add k-mask test, kortest intrinsics.
558 * config/i386/avx512dqintrin.h: Ditto.
559 * config/i386/avx512fintrin.h: Ditto.
560 * gcc/config/i386/i386.c: Handle new builtins.
561 * config/i386/i386-builtin.def: Add new builtins.
562 * config/i386/sse.md (ktest<mode>, kortest<mode>): New.
563 (UNSPEC_KORTEST, UNSPEC_KTEST): New.
565 2017-01-23 Jakub Jelinek <jakub@redhat.com>
566 Martin Liska <mliska@suse.cz>
568 * asan.h: Define ASAN_USE_AFTER_SCOPE_ATTRIBUTE.
569 * asan.c (asan_expand_poison_ifn): Support stores and use
570 appropriate ASAN report function.
571 * internal-fn.c (expand_ASAN_POISON_USE): New function.
572 * internal-fn.def (ASAN_POISON_USE): Declare.
573 * tree-into-ssa.c (maybe_add_asan_poison_write): New function.
574 (maybe_register_def): Create ASAN_POISON_USE when sanitizing.
575 * tree-ssa-dce.c (eliminate_unnecessary_stmts): Remove
576 ASAN_POISON calls w/o LHS.
577 * tree-ssa.c (execute_update_addresses_taken): Create clobber
578 for ASAN_MARK (UNPOISON, &x, ...) in order to prevent usage of a LHS
579 from ASAN_MARK (POISON, &x, ...) coming to a PHI node.
580 * gimplify.c (asan_poison_variables): Add attribute
581 use_after_scope_memory to variables that really needs to live
583 * tree-ssa.c (is_asan_mark_p): Do not rewrite into SSA when
584 having the attribute.
586 2017-01-23 Martin Liska <mliska@suse.cz>
588 * asan.c (create_asan_shadow_var): New function.
589 (asan_expand_poison_ifn): Likewise.
590 * asan.h (asan_expand_poison_ifn): New declaration.
591 * internal-fn.c (expand_ASAN_POISON): Likewise.
592 * internal-fn.def (ASAN_POISON): New builtin.
593 * sanopt.c (pass_sanopt::execute): Expand
594 asan_expand_poison_ifn.
595 * tree-inline.c (copy_decl_for_dup_finish): Make function
597 * tree-inline.h (copy_decl_for_dup_finish): Likewise.
598 * tree-ssa.c (is_asan_mark_p): New function.
599 (execute_update_addresses_taken): Rewrite local variables
600 (identified just by use-after-scope as addressable) into SSA.
602 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
604 * doc/install.texi (Specific): opensource.apple.com uses https
605 now. Remove trailing slash.
607 2017-01-22 Gerald Pfeifer <gerald@pfeifer.com>
609 * README.Portability: Remove note on an Irix compatibility issue.
611 2017-01-22 Dimitry Andric <dim@FreeBSD.org>
613 * gcov.c (INCLUDE_ALGORITHM): Define.
614 (INCLUDE_VECTOR): Define.
615 No longer include <vector> and <algorithm> directly.
617 2017-01-21 Gerald Pfeifer <gerald@pfeifer.com>
619 * doc/extend.texi (Thread-Local): Change www.akkadia.org reference
621 * doc/invoke.texi (Code Gen Options): Ditto.
623 2017-01-21 Jan Hubicka <hubicka@ucw.cz>
626 * cfg.c (update_bb_profile_for_threading): Fix updating of probablity.
628 2017-01-21 Bernd Schmidt <bschmidt@redhat.com>
630 rtl-optimization/79125
631 * cprop.c (local_cprop_pass): Handle cases where we make an
634 2017-01-20 Segher Boessenkool <segher@kernel.crashing.org>
638 * config/rs6000/rs6000.c (rs6000_gimplify_va_arg): Adjust address to
639 read from, for big endian.
641 2017-01-20 Jiong Wang <jiong.wang@arm.com>
643 * config/aarch64/aarch64-builtins.c (aarch64_init_builtins): Register
644 register pauth builtins for LP64 only.
646 2017-01-20 Marek Polacek <polacek@redhat.com>
649 * gimplify.c (should_warn_for_implicit_fallthrough): Handle consecutive
652 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
654 * omp-expand.c (expand_omp_simd): Clear PROP_gimple_lomp_dev regardless
656 * omp-offload.c (pass_omp_device_lower::gate): Use PROP_gimple_lomp_dev.
657 * passes.c (dump_properties): Handle PROP_gimple_lomp_dev.
658 * tree-inline.c (expand_call_inline): Propagate PROP_gimple_lomp_dev.
660 2017-01-20 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
663 * config/arm/arm.c (neon_valid_immediate): Reject vector constants
664 in big-endian mode when they are not a single duplicated value.
666 2017-01-20 Richard Biener <rguenther@suse.de>
668 * BASE-VER: Bump to 7.0.1.
670 2017-01-20 Alexander Monakov <amonakov@ispras.ru>
672 * omp-low.c (omplow_simd_context): New struct. Use it...
673 (lower_rec_simd_input_clauses): ...here and...
674 (lower_rec_input_clauses): ...here to hold common data. Adjust all
675 references to idx, lane, max_vf, is_simt.
677 2017-01-20 Graham Markall <graham.markall@embecosm.com>
679 * config/arc/arc.h (LINK_SPEC): Use arclinux_nps emulation when
682 2017-01-20 Martin Jambor <mjambor@suse.cz>
684 * hsa.h: Renaed to hsa-common.h. Adjusted a comment.
685 * hsa.c: Renaed to hsa-common.c. Change include of gt-hsa.h to
687 * Makefile.in (OBJS): Rename hsa.o to hsa-common.o.
688 (GTFILES): Rename hsa.c to hsa-common.c.
689 * hsa-brig.c: Change include of hsa.h to hsa-common.h.
690 * hsa-dump.c: Likewise.
691 * hsa-gen.c: Likewise.
692 * hsa-regalloc.c: Likewise.
693 * ipa-hsa.c: Likewise.
694 * omp-expand.c: Likewise.
695 * omp-low.c: Likewise.
696 * toplev.c: Likewise.
698 2017-01-20 Marek Polacek <polacek@redhat.com>
701 * doc/invoke.texi: Document -Wduplicated-branches.
702 * fold-const.c (operand_equal_p): Handle MODIFY_EXPR, INIT_EXPR,
703 COMPOUND_EXPR, PREDECREMENT_EXPR, PREINCREMENT_EXPR,
704 POSTDECREMENT_EXPR, POSTINCREMENT_EXPR, CLEANUP_POINT_EXPR, EXPR_STMT,
705 STATEMENT_LIST, and RETURN_EXPR. For non-pure non-const functions
706 return 0 only when not OEP_LEXICOGRAPHIC.
707 (fold_build_cleanup_point_expr): Use the expression
708 location when building CLEANUP_POINT_EXPR.
709 * tree-core.h (enum operand_equal_flag): Add OEP_LEXICOGRAPHIC.
710 * tree.c (add_expr): Handle error_mark_node.
712 2017-01-20 Martin Liska <mliska@suse.cz>
715 * tree-profile.c (init_ic_make_global_vars): Do not call
717 (gimple_init_gcov_profiler): Likewise.
719 2017-01-20 Martin Liska <mliska@suse.cz>
722 * cgraph.h (maybe_create_reference): Remove argument and
724 * cgraphclones.c (cgraph_node::create_virtual_clone): Remove one
726 * ipa-cp.c (create_specialized_node): Likewise.
727 * symtab.c (symtab_node::maybe_create_reference): Handle
728 VAR_DECLs and ADDR_EXPRs and select ipa_ref_use type.
730 2017-01-20 Martin Liska <mliska@suse.cz>
732 * read-rtl-function.c (function_reader::create_function): Use
733 build_decl instread of build_decl_stat.
735 2017-01-20 Andrew Senkevich <andrew.senkevich@intel.com>
737 * config/i386/avx512bwintrin.h: Add k-mask registers shift intrinsics.
738 * config/i386/avx512dqintrin.h: Ditto.
739 * config/i386/avx512fintrin.h: Ditto.
740 * config/i386/i386-builtin-types.def: Add new types.
741 * gcc/config/i386/i386.c: Handle new types.
742 * config/i386/i386-builtin.def (__builtin_ia32_kshiftliqi)
743 (__builtin_ia32_kshiftlihi, __builtin_ia32_kshiftlisi)
744 (__builtin_ia32_kshiftlidi, __builtin_ia32_kshiftriqi)
745 (__builtin_ia32_kshiftrihi, __builtin_ia32_kshiftrisi)
746 (__builtin_ia32_kshiftridi): New.
747 * config/i386/sse.md (k<code><mode>): Rename *k<code><mode>.
749 2017-01-19 Segher Boessenkool <segher@kernel.crashing.org>
753 * config/rs6000/rs6000.c (TARGET_STACK_PROTECT_GUARD): Unconditionally
754 define to rs6000_init_stack_protect_guard.
755 (rs6000_init_stack_protect_guard): New function.
757 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
758 Yunqiang Su <yunqiang.su@imgtec.com>
760 * config.gcc (supported_defaults): Add madd4.
761 (with_madd4): Add validation.
762 (all_defaults): Add madd4.
763 * config/mips/mips.opt (mmadd4): New option.
764 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
766 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
768 (ISA_HAS_UNFUSED_MADD4): Gate with mips_madd4.
769 (ISA_HAS_FUSED_MADD4): Likewise.
770 * gcc/doc/invoke.texi (-mmadd4): Document the new option.
771 * gcc/doc/install.texi (--with-madd4): Document the new option.
773 2017-01-19 Jiong Wang <jiong.wang@arm.com>
775 * config/aarch64/aarch64-builtins.c (enum aarch64_builtins): New
776 entries for AARCH64_PAUTH_BUILTIN_XPACLRI,
777 AARCH64_PAUTH_BUILTIN_PACIA1716, AARCH64_PAUTH_BUILTIN_AUTIA1716.
778 (aarch64_init_pauth_hint_builtins): New.
779 (aarch64_init_builtins): Call aarch64_init_pauth_hint_builtins.
780 (aarch64_expand_builtin): Expand new builtins.
782 2017-01-19 Jiong Wang <jiong.wang@arm.com>
784 * reg-notes.def (CFA_TOGGLE_RA_MANGLE): New reg-note.
785 * combine-stack-adj.c (no_unhandled_cfa): Handle
786 REG_CFA_TOGGLE_RA_MANGLE.
787 * dwarf2cfi.c (dwarf2out_frame_debug): Handle REG_CFA_TOGGLE_RA_MANGLE.
788 * config/aarch64/aarch64.c (aarch64_expand_prologue): Generates DWARF
789 info for return address signing.
790 (aarch64_expand_epilogue): Likewise.
792 2017-01-19 Jiong Wang <jiong.wang@arm.com>
794 * config/aarch64/aarch64-opts.h (aarch64_function_type): New enum.
795 * config/aarch64/aarch64-protos.h
796 (aarch64_return_address_signing_enabled): New declaration.
797 * config/aarch64/aarch64.c (aarch64_return_address_signing_enabled):
799 (aarch64_expand_prologue): Sign return address before it's pushed onto
801 (aarch64_expand_epilogue): Authenticate return address fetched from
803 (aarch64_override_options): Sanity check for ILP32 and ISA level.
804 (aarch64_attributes): New function attributes for "sign-return-address".
805 * config/aarch64/aarch64.md (UNSPEC_AUTI1716, UNSPEC_AUTISP,
806 UNSPEC_PACI1716, UNSPEC_PACISP, UNSPEC_XPACLRI): New unspecs.
807 ("*do_return"): Generate combined instructions according to key index.
808 ("<pauth_mnem_prefix>sp", "<pauth_mnem_prefix1716", "xpaclri"): New.
809 * config/aarch64/iterators.md (PAUTH_LR_SP, PAUTH_17_16): New integer
811 (pauth_mnem_prefix, pauth_hint_num_a): New integer attributes.
812 * config/aarch64/aarch64.opt (msign-return-address=): New.
813 * doc/extend.texi (AArch64 Function Attributes): Documents
814 "sign-return-address=".
815 * doc/invoke.texi (AArch64 Options): Documents "-msign-return-address=".
817 2017-01-19 Matthew Fortune <matthew.fortune@imgtec.com>
819 * doc/invoke.texi: Add missing -mlxc1-sxc1 options to
820 overall option summary.
822 2017-01-19 Jiong Wang <jiong.wang@arm.com>
824 * config/aarch64/aarch64-arches.def: New entry for "armv8.3-a".
825 * config/aarch64/aarch64.h (AARCH64_FL_V8_3, AARCH64_FL_FOR_ARCH8_3,
826 AARCH64_ISA_V8_3, TARGET_ARMV8_3): New.
827 * doc/invoke.texi (AArch64 Options): Document "armv8.3-a".
829 2017-01-19 Michael Meissner <meissner@linux.vnet.ibm.com>
831 * config/rs6000/rs6000-cpus.def (ISA_3_0_MASKS_SERVER): Enable
832 -mpower9-minmax by default for -mcpu=power9.
833 (ISA_3_MASKS_IEEE): Require -mvsx-small-integer to enable IEEE
834 128-bit floating point.
836 2017-01-20 Alan Modra <amodra@gmail.com>
838 * config/rs6000/rs6000.md (cmpstrnsi, cmpstrsi): Fail if
841 2017-01-20 Alan Modra <amodra@gmail.com>
844 * config/rs6000/rs6000.c (expand_strn_compare): Get the asm name
845 for strcmp and strncmp from corresponding builtin decl.
847 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
849 * config.gcc (x86_64-*-rtems*): Use i386/rtemself.h
850 instead of i386/rtems-64.h.
851 * config/i386/rtems-64.h: Remove.
853 2017-01-19 Uros Bizjak <ubizjak@gmail.com>
857 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
859 * config/i386/rtemself.h (LONG_DOUBLE_TYPE_SIZE): New define.
861 2017-01-19 Tamar Christina <tamar.christina@arm.com>
863 * config/aarch64/aarch64.c (aarch64_simd_gen_const_vector_dup):
864 Change int to HOST_WIDE_INT.
865 * config/aarch64/aarch64-protos.h
866 (aarch64_simd_gen_const_vector_dup): Likewise.
867 * config/aarch64/aarch64-simd.md: Add copysign<mode>3.
869 2017-01-19 David Malcolm <dmalcolm@redhat.com>
871 * langhooks-def.h (lhd_type_for_size): New decl.
872 (LANG_HOOKS_TYPE_FOR_SIZE): Define as lhd_type_for_size.
873 * langhooks.c (lhd_type_for_size): New function, taken from
876 2017-01-19 Pat Haugen <pthaugen@us.ibm.com>
878 * config/rs6000/power9.md (power9-alu): Remove 'cmp' type and add
879 define_bypass for CR latency.
880 (power9-cracked-alu): Update bypass latency and remove power9-branch.
881 (power9-alu2): Add define_bypass for CR latency.
883 (power9-mul): Update insn latency.
884 (power9-mul-compare): Update insn latency, bypass latency and remove
887 2016-01-19 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
889 * config/aarch64/aarch64-protos.h (aarch64_nopcrelative_literal_loads):
891 * config/aarch64/aarch64.md
892 (aarch64_reload_movcp<GPF_TF:mode><P:mode>): Delete reference to
893 aarch64_nopcrelative_literal_loads.
894 (aarch64_reload_movcp<VALL:mode><P:mode>): Likewise.
896 2017-01-19 Chenghua Xu <paul.hua.gm@gmail.com>
898 * config/mips/mips.h (ISA_HAS_FUSED_MADD4): Enable for
900 (ISA_HAS_UNFUSED_MADD4): Exclude TARGET_LOONGSON_3A.
902 2017-01-19 Doug Gilmore <doug.gilmore@imgtec.com>
905 * config.gcc (supported_defaults): Add lxc1-sxc1.
906 (with_lxc1_sxc1): Add validation.
907 (all_defaults): Add lxc1-sxc1.
908 * config/mips/mips.opt (mlxc1-sxc1): New option.
909 * gcc/config/mips/mips.h (OPTION_DEFAULT_SPECS): Add a default for
911 (TARGET_CPU_CPP_BUILTINS): Add builtin_define for
913 (ISA_HAS_LXC1_SXC1): Gate with mips_lxc1_sxc1.
914 * gcc/doc/invoke.texi (-mlxc1-sxc1): Document the new option.
915 * doc/install.texi (--with-lxc1-sxc1): Document the new option.
917 2017-01-19 Richard Biener <rguenther@suse.de>
919 PR tree-optimization/72488
920 * tree-ssa-sccvn.c (run_scc_vn): When we abort the VN make
921 sure to restore SSA info.
922 * tree-ssa.c (verify_ssa): Verify SSA info is not shared.
924 2017-01-19 Richard Earnshaw <rearnsha@arm.com>
926 PR rtl-optimization/79121
927 * expr.c (expand_expr_real_2, case LSHIFT_EXPR): Look at the signedness
928 of the inner type when shifting an extended value.
930 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
933 * symtab.c (symtab_node::equal_address_to): Fix comparing of
934 interposable aliases.
936 2017-01-18 Peter Bergner <bergner@vnet.ibm.com>
939 * config/rs6000/spe.md (mov_si<mode>_e500_subreg0): Fix constraints.
940 Use the evmergelohi instruction.
941 (mov_si<mode>_e500_subreg4_2_le): Likewise.
942 (mov_sitf_e500_subreg8_2_be): Likewise.
943 (mov_sitf_e500_subreg12_2_le): Likewise.
944 (mov_si<mode>_e500_subreg0_2_le): Fix constraints.
945 (mov_si<mode>_e500_subreg4_2_be): Likewise.
946 (mov_sitf_e500_subreg8_2_le): Likewise.
947 (mov_sitf_e500_subreg12_2_be): Likewise.
949 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
951 * config/rs6000/altivec.md (altivec_vbpermq): Change "type"
952 attribute from vecsimple to vecperm.
953 (altivec_vbpermq2): Likewise.
955 2017-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
958 * config/rs6000/altivec.h: Fix typo of vec_cntlz to vec_cnttz.
960 2017-01-18 Aaron Sawdey <acsawdey@linux.vnet.ibm.com>
961 * config/rs6000/rs6000-protos.h (expand_strn_compare): Add arg.
962 * config/rs6000/rs6000.c (expand_strn_compare): Add ability to expand
963 strcmp. Fix bug where comparison didn't stop with zero byte. Fix
964 case where N arg is SIZE_MAX.
965 * config/rs6000/rs6000.md (cmpstrnsi): Args to expand_strn_compare.
966 (cmpstrsi): Add pattern.
968 2017-01-18 Michael Meissner <meissner@linux.vnet.ibm.com>
970 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
971 __builtin_vec_revb builtins.
972 * config/rs6000/rs6000-builtins.def (P9V_BUILTIN_XXBRQ_V16QI): Add
973 built-in functions to support generation of the ISA 3.0 XXBR<x>
974 vector byte reverse instructions.
975 (P9V_BUILTIN_XXBRQ_V1TI): Likewise.
976 (P9V_BUILTIN_XXBRD_V2DI): Likewise.
977 (P9V_BUILTIN_XXBRD_V2DF): Likewise.
978 (P9V_BUILTIN_XXBGW_V4SI): Likewise.
979 (P9V_BUILTIN_XXBGW_V4SF): Likewise.
980 (P9V_BUILTIN_XXBGH_V8HI): Likewise.
981 (P9V_BUILTIN_VEC_REVB): Likewise.
982 * config/rs6000/vsx.md (p9_xxbrq_v1ti): New insns/expanders to
983 generate the ISA 3.0 XXBR<x> vector byte reverse instructions.
984 (p9_xxbrq_v16qi): Likewise.
985 (p9_xxbrd_<mode>, VSX_D iterator): Likewise.
986 (p9_xxbrw_<mode>, VSX_W iterator): Likewise.
987 (p9_xxbrh_v8hi): Likewise.
988 * config/rs6000/altivec.h (vec_revb): Define if ISA 3.0.
989 * doc/extend.texi (RS/6000 Altivec Built-ins): Document the
990 vec_revb built-in functions.
992 2017-01-18 Uros Bizjak <ubizjak@gmail.com>
994 PR rtl-optimization/78952
995 * config/i386/i386.md (any_extract): New code iterator.
996 (*insvqi_2): Use any_extract for source operand.
997 (*insvqi_3): Use any_shiftrt for source operand.
999 2017-01-18 Wilco Dijkstra <wdijkstr@arm.com>
1001 * config/aarch64/aarch64.c (aarch64_sched_adjust_priority)
1003 (TARGET_SCHED_ADJUST_PRIORITY): Define target hook.
1005 2017-01-18 Matthias Klose <doko@ubuntu.com>
1007 * doc/install.texi: Allow default for --with-target-bdw-gc-include.
1009 2016-01-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1011 * config/rs6000/altivec.h (vec_bperm): Change #define.
1012 * config/rs6000/altivec.md (UNSPEC_VBPERMD): New enum constant.
1013 (altivec_vbpermq2): New define_insn.
1014 (altivec_vbpermd): Likewise.
1015 * config/rs6000/rs6000-builtin.def (VBPERMQ2): New monomorphic
1017 (VBPERMD): Likewise.
1018 (VBPERM): New polymorphic function interface.
1019 * config/rs6000/r6000-c.c (altivec_overloaded_builtins_table):
1020 Add entries for P9V_BUILTIN_VEC_VBPERM.
1021 * doc/extend.texi: Add interfaces for vec_bperm.
1023 2017-01-18 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
1025 * config/s390/s390-c.c (s390_expand_overloaded_builtin): Downcase
1026 first letter of error messages.
1027 (s390_resolve_overloaded_builtin): Likewise.
1028 * config/s390/s390.c (s390_expand_builtin): Likewise.
1029 (s390_invalid_arg_for_unprototyped_fn): Likewise.
1030 (s390_valid_target_attribute_inner_p): Likewise.
1031 * config/s390/s390.md ("tabort"): Likewise.
1033 2017-01-18 Toma Tabacu <toma.tabacu@imgtec.com>
1035 * config/mips/mips.h (ISA_HAS_DIV3): Remove unused macro.
1036 (ISA_AVOID_DIV_HILO): New macro.
1037 (ISA_HAS_DIV): Use new ISA_AVOID_DIV_HILO macro.
1038 (ISA_HAS_DDIV): Likewise.
1040 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1042 * doc/invoke.texi (fabi-version): Correct number of occurrences.
1044 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1046 * doc/invoke.texi (fabi-version): Spelling fix.
1048 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1051 * doc/invoke.texi (fabi-version): Mention mangling fix for
1054 2017-01-18 Markus Trippelsdorf <markus@trippelsdorf.de>
1057 * doc/invoke.texi (fabi-version): Document discriminator mangling.
1059 2017-01-17 Segher Boessenkool <segher@kernel.crashing.org>
1062 * config/rs6000/rs6000-opts.h (stack_protector_guard): New enum.
1063 * config/rs6000/rs6000.c (rs6000_option_override_internal): Handle
1065 * config/rs6000/rs6000.md (stack_protect_set): Handle the new more
1067 (stack_protect_test): Ditto.
1068 * config/rs6000/rs6000.opt (mstack-protector-guard=,
1069 mstack-protector-guard-reg=, mstack-protector-guard-offset=): New
1071 * doc/invoke.texi (Option Summary) [RS/6000 and PowerPC Options]:
1072 Add -mstack-protector-guard=, -mstack-protector-guard-reg=, and
1073 -mstack-protector-guard-offset=.
1074 (RS/6000 and PowerPC Options): Ditto.
1076 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1078 * config/i386/i386.h (MASK_CLASS_P): New define.
1079 * config/i386/i386.c (inline_secondary_memory_needed): Ensure that
1080 there are no registers from different register sets also when
1081 mask registers are used. Update function comment.
1082 * config/i386/i386.md (*movsi_internal): Split (*k/*krm) alternative
1083 to (*k/*r) and (*k/*km) alternatives.
1085 2017-01-17 Wilco Dijkstra <wdijkstr@arm.com>
1087 * config/aarch64/aarch64.md (eh_return): Remove pattern and splitter.
1088 * config/aarch64/aarch64.h (AARCH64_EH_STACKADJ_REGNUM): Remove.
1089 (EH_RETURN_HANDLER_RTX): New define.
1090 * config/aarch64/aarch64.c (aarch64_frame_pointer_required):
1091 Force frame pointer in EH return functions.
1092 (aarch64_expand_epilogue): Add barrier for eh_return.
1093 (aarch64_final_eh_return_addr): Remove.
1094 (aarch64_eh_return_handler_rtx): New function.
1095 * config/aarch64/aarch64-protos.h (aarch64_final_eh_return_addr):
1097 (aarch64_eh_return_handler_rtx): New prototype.
1099 2017-01-17 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1101 * config/rs6000/altivec.h (vec_rlmi): New #define.
1102 (vec_vrlnm): Likewise.
1103 (vec_rlnm): Likewise.
1104 * config/rs6000/altivec.md (UNSPEC_VRLMI): New UNSPEC enum value.
1105 (UNSPEC_VRLNM): Likewise.
1106 (VIlong): New mode iterator.
1107 (altivec_vrl<VI_char>mi): New define_insn.
1108 (altivec_vrl<VI_char>nm): Likewise.
1109 * config/rs6000/rs6000-builtin.def (VRLWNM): New monomorphic
1112 (RLNM): New polymorphic function entry.
1113 (VRLWMI): New monomorphic function entry.
1115 (RLMI): New polymorphic function entry.
1116 * config/rs6000/r6000-c.c (altivec_overloaded_builtin_table): Add
1117 new entries for P9V_BUILTIN_VEC_RLMI and P9V_BUILTIN_VEC_RLNM.
1118 * doc/extend.texi: Add description of vec_rlmi, vec_rlnm, and
1121 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1124 * dwarf2out.c (field_byte_offset): Restore the
1125 PCC_BITFIELD_TYPE_MATTERS behavior for INTEGER_CST DECL_FIELD_OFFSET
1126 and DECL_FIELD_BIT_OFFSET. Use fold_build2 instead of build2 + fold.
1127 (analyze_variants_discr, gen_variant_part): Use fold_build2 instead
1130 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1133 * config/aarch64/aarch64.c (TARGET_CUSTOM_FUNCTION_DESCRIPTORS): Define
1135 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1138 * dwarf2out.c (add_data_member_location_attribute): For constant
1139 offset bitfield emit for -gdwarf-5 DW_AT_data_bit_offset attribute
1140 instead of DW_AT_data_member_location, DW_AT_bit_offset and
1141 DW_AT_byte_size attributes.
1143 2017-01-17 Eric Botcazou <ebotcazou@adacore.com>
1145 * config/rs6000/rs6000.c (rs6000_emit_move): Also use a TOC reference
1146 after forcing to constant memory when the code model is medium.
1148 2017-01-17 Julia Koval <julia.koval@intel.com>
1151 * config/i386/avx512fintrin.h
1152 (_mm512_i32gather_ps): Change __addr type to void const*.
1153 (_mm512_mask_i32gather_ps): Ditto.
1154 (_mm512_i32gather_pd): Ditto.
1155 (_mm512_mask_i32gather_pd): Ditto.
1156 (_mm512_i64gather_ps): Ditto.
1157 (_mm512_mask_i64gather_ps): Ditto.
1158 (_mm512_i64gather_pd): Ditto.
1159 (_mm512_mask_i64gather_pd): Ditto.
1160 (_mm512_i32gather_epi32): Ditto.
1161 (_mm512_mask_i32gather_epi32): Ditto.
1162 (_mm512_i32gather_epi64): Ditto.
1163 (_mm512_mask_i32gather_epi64): Ditto.
1164 (_mm512_i64gather_epi32): Ditto.
1165 (_mm512_mask_i64gather_epi32): Ditto.
1166 (_mm512_i64gather_epi64): Ditto.
1167 (_mm512_mask_i64gather_epi64): Ditto.
1168 (_mm512_i32scatter_ps): Change __addr type to void*.
1169 (_mm512_mask_i32scatter_ps): Ditto.
1170 (_mm512_i32scatter_pd): Ditto.
1171 (_mm512_mask_i32scatter_pd): Ditto.
1172 (_mm512_i64scatter_ps): Ditto.
1173 (_mm512_mask_i64scatter_ps): Ditto.
1174 (_mm512_i64scatter_pd): Ditto.
1175 (_mm512_mask_i64scatter_pd): Ditto.
1176 (_mm512_i32scatter_epi32): Ditto.
1177 (_mm512_mask_i32scatter_epi32): Ditto.
1178 (_mm512_i32scatter_epi64): Ditto.
1179 (_mm512_mask_i32scatter_epi64): Ditto.
1180 (_mm512_i64scatter_epi32): Ditto.
1181 (_mm512_mask_i64scatter_epi32): Ditto.
1182 (_mm512_i64scatter_epi64): Ditto.
1183 (_mm512_mask_i64scatter_epi64): Ditto.
1184 * config/i386/avx512pfintrin.h
1185 (_mm512_mask_prefetch_i32gather_pd): Change __addr type to void const*.
1186 (_mm512_mask_prefetch_i32gather_ps): Ditto.
1187 (_mm512_mask_prefetch_i64gather_pd): Ditto.
1188 (_mm512_mask_prefetch_i64gather_ps): Ditto.
1189 (_mm512_prefetch_i32scatter_pd): Change __addr type to void*.
1190 (_mm512_prefetch_i32scatter_ps): Ditto.
1191 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
1192 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
1193 (_mm512_prefetch_i64scatter_pd): Ditto.
1194 (_mm512_prefetch_i64scatter_ps): Ditto.
1195 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
1196 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
1197 * config/i386/avx512vlintrin.h
1198 (_mm256_mmask_i32gather_ps): Change __addr type to void const*.
1199 (_mm_mmask_i32gather_ps): Ditto.
1200 (_mm256_mmask_i32gather_pd): Ditto.
1201 (_mm_mmask_i32gather_pd): Ditto.
1202 (_mm256_mmask_i64gather_ps): Ditto.
1203 (_mm_mmask_i64gather_ps): Ditto.
1204 (_mm256_mmask_i64gather_pd): Ditto.
1205 (_mm_mmask_i64gather_pd): Ditto.
1206 (_mm256_mmask_i32gather_epi32): Ditto.
1207 (_mm_mmask_i32gather_epi32): Ditto.
1208 (_mm256_mmask_i32gather_epi64): Ditto.
1209 (_mm_mmask_i32gather_epi64): Ditto.
1210 (_mm256_mmask_i64gather_epi32): Ditto.
1211 (_mm_mmask_i64gather_epi32): Ditto.
1212 (_mm256_mmask_i64gather_epi64): Ditto.
1213 (_mm_mmask_i64gather_epi64): Ditto.
1214 (_mm256_i32scatter_ps): Change __addr type to void*.
1215 (_mm256_mask_i32scatter_ps): Ditto.
1216 (_mm_i32scatter_ps): Ditto.
1217 (_mm_mask_i32scatter_ps): Ditto.
1218 (_mm256_i32scatter_pd): Ditto.
1219 (_mm256_mask_i32scatter_pd): Ditto.
1220 (_mm_i32scatter_pd): Ditto.
1221 (_mm_mask_i32scatter_pd): Ditto.
1222 (_mm256_i64scatter_ps): Ditto.
1223 (_mm256_mask_i64scatter_ps): Ditto.
1224 (_mm_i64scatter_ps): Ditto.
1225 (_mm_mask_i64scatter_ps): Ditto.
1226 (_mm256_i64scatter_pd): Ditto.
1227 (_mm256_mask_i64scatter_pd): Ditto.
1228 (_mm_i64scatter_pd): Ditto.
1229 (_mm_mask_i64scatter_pd): Ditto.
1230 (_mm256_i32scatter_epi32): Ditto.
1231 (_mm256_mask_i32scatter_epi32): Ditto.
1232 (_mm_i32scatter_epi32): Ditto.
1233 (_mm_mask_i32scatter_epi32): Ditto.
1234 (_mm256_i32scatter_epi64): Ditto.
1235 (_mm256_mask_i32scatter_epi64): Ditto.
1236 (_mm_i32scatter_epi64): Ditto.
1237 (_mm_mask_i32scatter_epi64): Ditto.
1238 (_mm256_i64scatter_epi32): Ditto.
1239 (_mm256_mask_i64scatter_epi32): Ditto.
1240 (_mm_i64scatter_epi32): Ditto.
1241 (_mm_mask_i64scatter_epi32): Ditto.
1242 (_mm256_i64scatter_epi64): Ditto.
1243 (_mm256_mask_i64scatter_epi64): Ditto.
1244 (_mm_i64scatter_epi64): Ditto.
1245 (_mm_mask_i64scatter_epi64): Ditto.
1246 * config/i386/i386-builtin-types.def (V16SF_V16SF_PCFLOAT_V16SI_HI_INT)
1247 (V8DF_V8DF_PCDOUBLE_V8SI_QI_INT, V8SF_V8SF_PCFLOAT_V8DI_QI_INT)
1248 (V8DF_V8DF_PCDOUBLE_V8DI_QI_INT, V16SI_V16SI_PCINT_V16SI_HI_INT)
1249 (V8DI_V8DI_PCINT64_V8SI_QI_INT, V8SI_V8SI_PCINT_V8DI_QI_INT)
1250 (V8DI_V8DI_PCINT64_V8DI_QI_INT, V2DF_V2DF_PCDOUBLE_V4SI_QI_INT)
1251 (V4DF_V4DF_PCDOUBLE_V4SI_QI_INT, V2DF_V2DF_PCDOUBLE_V2DI_QI_INT)
1252 (V4DF_V4DF_PCDOUBLE_V4DI_QI_INT, V4SF_V4SF_PCFLOAT_V4SI_QI_INT)
1253 (V8SF_V8SF_PCFLOAT_V8SI_QI_INT, V4SF_V4SF_PCFLOAT_V2DI_QI_INT)
1254 (V4SF_V4SF_PCFLOAT_V4DI_QI_INT, V2DI_V2DI_PCINT64_V4SI_QI_INT)
1255 (V4DI_V4DI_PCINT64_V4SI_QI_INT, V2DI_V2DI_PCINT64_V2DI_QI_INT)
1256 (V4DI_V4DI_PCINT64_V4DI_QI_INT, V4SI_V4SI_PCINT_V4SI_QI_INT)
1257 (V8SI_V8SI_PCINT_V8SI_QI_INT, V4SI_V4SI_PCINT_V2DI_QI_INT)
1258 (V4SI_V4SI_PCINT_V4DI_QI_INT, VOID_PFLOAT_HI_V16SI_V16SF_INT)
1259 (VOID_PFLOAT_QI_V8SI_V8SF_INT, VOID_PFLOAT_QI_V4SI_V4SF_INT)
1260 (VOID_PDOUBLE_QI_V8SI_V8DF_INT, VOID_PDOUBLE_QI_V4SI_V4DF_INT)
1261 (VOID_PDOUBLE_QI_V4SI_V2DF_INT, VOID_PFLOAT_QI_V8DI_V8SF_INT)
1262 (VOID_PFLOAT_QI_V4DI_V4SF_INT, VOID_PFLOAT_QI_V2DI_V4SF_INT)
1263 (VOID_PDOUBLE_QI_V8DI_V8DF_INT, VOID_PDOUBLE_QI_V4DI_V4DF_INT)
1264 (VOID_PDOUBLE_QI_V2DI_V2DF_INT, VOID_PINT_HI_V16SI_V16SI_INT)
1265 (VOID_PINT_QI_V8SI_V8SI_INT, VOID_PINT_QI_V4SI_V4SI_INT)
1266 (VOID_PLONGLONG_QI_V8SI_V8DI_INT, VOID_PLONGLONG_QI_V4SI_V4DI_INT)
1267 (VOID_PLONGLONG_QI_V4SI_V2DI_INT, VOID_PINT_QI_V8DI_V8SI_INT)
1268 (VOID_PINT_QI_V4DI_V4SI_INT, VOID_PINT_QI_V2DI_V4SI_INT)
1269 (VOID_PLONGLONG_QI_V8DI_V8DI_INT, VOID_QI_V8SI_PCINT64_INT_INT)
1270 (VOID_PLONGLONG_QI_V4DI_V4DI_INT, VOID_PLONGLONG_QI_V2DI_V2DI_INT)
1271 (VOID_HI_V16SI_PCINT_INT_INT, VOID_QI_V8DI_PCINT64_INT_INT)
1272 (VOID_QI_V8DI_PCINT_INT_INT): Remove.
1273 (V16SF_V16SF_PCVOID_V16SI_HI_INT, V8DF_V8DF_PCVOID_V8SI_QI_INT)
1274 (V8SF_V8SF_PCVOID_V8DI_QI_INT, V8DF_V8DF_PCVOID_V8DI_QI_INT)
1275 (V16SI_V16SI_PCVOID_V16SI_HI_INT, V8DI_V8DI_PCVOID_V8SI_QI_INT)
1276 (V8SI_V8SI_PCVOID_V8DI_QI_INT, V8DI_V8DI_PCVOID_V8DI_QI_INT)
1277 (VOID_PVOID_HI_V16SI_V16SF_INT, VOID_PVOID_QI_V8SI_V8DF_INT)
1278 (VOID_PVOID_QI_V8DI_V8SF_INT, VOID_PVOID_QI_V8DI_V8DF_INT)
1279 (VOID_PVOID_HI_V16SI_V16SI_INT, VOID_PVOID_QI_V8SI_V8DI_INT)
1280 (VOID_PVOID_QI_V8DI_V8SI_INT, VOID_PVOID_QI_V8DI_V8DI_INT)
1281 (V2DF_V2DF_PCVOID_V4SI_QI_INT, V4DF_V4DF_PCVOID_V4SI_QI_INT)
1282 (V2DF_V2DF_PCVOID_V2DI_QI_INT, V4DF_V4DF_PCVOID_V4DI_QI_INT
1283 (V4SF_V4SF_PCVOID_V4SI_QI_INT, V8SF_V8SF_PCVOID_V8SI_QI_INT)
1284 (V4SF_V4SF_PCVOID_V2DI_QI_INT, V4SF_V4SF_PCVOID_V4DI_QI_INT)
1285 (V2DI_V2DI_PCVOID_V4SI_QI_INT, V4DI_V4DI_PCVOID_V4SI_QI_INT)
1286 (V2DI_V2DI_PCVOID_V2DI_QI_INT, V4DI_V4DI_PCVOID_V4DI_QI_INT)
1287 (V4SI_V4SI_PCVOID_V4SI_QI_INT, V8SI_V8SI_PCVOID_V8SI_QI_INT)
1288 (V4SI_V4SI_PCVOID_V2DI_QI_INT, V4SI_V4SI_PCVOID_V4DI_QI_INT)
1289 (VOID_PVOID_QI_V8SI_V8SF_INT, VOID_PVOID_QI_V4SI_V4SF_INT)
1290 (VOID_PVOID_QI_V4SI_V4DF_INT, VOID_PVOID_QI_V4SI_V2DF_INT)
1291 (VOID_PVOID_QI_V4DI_V4SF_INT, VOID_PVOID_QI_V2DI_V4SF_INT)
1292 (VOID_PVOID_QI_V4DI_V4DF_INT, VOID_PVOID_QI_V2DI_V2DF_INT)
1293 (VOID_PVOID_QI_V8SI_V8SI_INT, VOID_PVOID_QI_V4SI_V4SI_INT)
1294 (VOID_PVOID_QI_V4SI_V4DI_INT, VOID_PVOID_QI_V4SI_V2DI_INT)
1295 (VOID_PVOID_QI_V4DI_V4SI_INT, VOID_PVOID_QI_V2DI_V4SI_INT)
1296 (VOID_PVOID_QI_V4DI_V4DI_INT, VOID_PVOID_QI_V2DI_V2DI_INT)
1297 (VOID_QI_V8SI_PCVOID_INT_INT, VOID_HI_V16SI_PCVOID_INT_INT)
1298 (VOID_QI_V8DI_PCVOID_INT_INT): Add.
1299 * config/i386/i386.c (ix86_init_mmx_sse_builtins): Adjust builtin
1300 definitions accordingly.
1302 2017-01-17 Kito Cheng <kito.cheng@gmail.com>
1303 Kuan-Lin Chen <kuanlinchentw@gmail.com>
1306 * internal-fn.c (expand_mul_overflow): Use convert_modes instead of
1309 2017-01-17 Vladimir Makarov <vmakarov@redhat.com>
1312 * ira-conflicts.c (ira_build_conflicts): Update total conflict
1313 hard regs for inner regno.
1315 2017-01-17 Martin Liska <mliska@suse.cz>
1318 * ipa-polymorphic-call.c (contains_type_p): Fix wrong
1319 assumption and add comment.
1321 2017-01-17 Nathan Sidwell <nathan@acm.org>
1323 * ipa-visibility.c (localize_node): New function, broken out of ...
1324 (function_and_variable_visibility): ... here. Call it.
1326 2017-01-17 Jan Hubicka <hubicka@ucw.cz>
1329 * tree-ssa-threadupdate.c (remove_ctrl_stmt_and_useless_edges):
1330 correctly set frequency of oudgoing edge.
1331 (duplicate_thread_path): Fix profile updating.
1333 2017-01-17 Jakub Jelinek <jakub@redhat.com>
1336 * configure.ac: Add GCC_BASE_VER.
1337 * Makefile.in (version): Use @get_gcc_base_ver@ instead of cat to get
1338 version from BASE-VER file.
1339 (CFLAGS-gcc.o): Add -DBASEVER=$(BASEVER_s).
1340 (gcc.o): Depend on $(BASEVER).
1341 * common.opt (dumpfullversion): New option.
1342 * gcc.c (driver_handle_option): Handle OPT_dumpfullversion.
1343 * doc/invoke.texi: Document -dumpfullversion.
1344 * doc/install.texi: Document --with-gcc-major-version-only.
1345 * configure: Regenerated.
1347 2017-01-17 Richard Biener <rguenther@suse.de>
1349 PR tree-optimization/71433
1350 * tree-vrp.c (register_new_assert_for): Merge same asserts
1351 on all incoming edges.
1352 (process_assert_insertions_for): Handle insertions at the
1355 2017-01-17 Gerald Pfeifer <gerald@pfeifer.com>
1357 * config/i386/cygwin.h (LIBGCJ_SONAME): Remove.
1358 * config/i386/mingw32.h (LIBGCJ_SONAME): Remove.
1360 2017-01-17 Kaz Kojima <kkojima@gcc.gnu.org>
1363 * config/sh/sh.md (cmpeqsi_t+1): Call copy_rtx to avoid invalid
1366 2017-01-17 Alan Modra <amodra@gmail.com>
1369 * config/rs6000/rs6000.md (elf_high, elf_low): Disable when pic.
1370 * config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Don't allow
1371 symbolic stack limit when pic.
1373 2017-01-16 Martin Sebor <msebor@redhat.com>
1375 PR tree-optimization/78608
1376 * gimple-ssa-sprintf.c (tree_digits): Avoid negating TYPE_MIN.
1378 2017-01-16 Jeff Law <law@redhat.com>
1381 2016-12-02 Tadek Kijkowski <tkijkowski@gmail.com>
1382 * Makefile.in (PREPROCESSOR_DEFINES): Add a level of indirection
1383 for several include directories that may be relative to sysroot.
1384 * config/i386/x-mingw32 (gplus_includedir): Define.
1385 (gplus_tool_includedir, gplus_backward_include_dir): Likewise.
1386 (native_system_includedir): Likewise.
1387 * config/i386/mingw32.h (STANDARD_STARTFILE_PREFIX_1): Do not
1388 override if TARGET_SYSTEM_ROOT is defined.
1389 (NATIVE_SYSTEM_HEADER_DIR): Likewise.
1391 PR tree-optimization/79090
1392 PR tree-optimization/33562
1393 PR tree-optimization/61912
1394 PR tree-optimization/77485
1395 * tree-ssa-dse.c (compute_trims): Accept STMT argument. Dump STMT
1396 and computed trims into the dump file.
1398 2017-01-17 Uros Bizjak <ubizjak@gmail.com>
1400 * config/i386/i386.h (LIMIT_RELOAD_CLASS): Remove.
1402 2017-01-16 Jakub Jelinek <jakub@redhat.com>
1405 * gimplify.c (gimplify_init_constructor): If want_value and
1406 object == lhs, unshare lhs to avoid invalid tree sharing. Formatting
1410 * loop-doloop.c (doloop_modify): Call unshare_all_rtl_in_chain on
1411 sequence. Formatting fixes.
1412 (doloop_optimize): Formatting fixes.
1415 * gcc.c (debug_level_greater_than_spec_func): New function.
1416 (static_spec_functions): Add debug-level-gt spec function.
1417 (ASM_DEBUG_SPEC, cpp_options): Use %:debug-level-gt(0) instead of
1419 * config/darwin.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1420 * config/darwin9.h (DSYMUTIL_SPEC, ASM_DEBUG_SPEC): Likewise.
1421 * common.opt (g, gcoff, gdwarf, gdwarf-, ggdb, gno-pubnames,
1422 gpubnames, ggnu-pubnames, gno-record-gcc-switches,
1423 grecord-gcc-switches, gno-strict-dwarf, gstrict-dwarf, gstabs,
1424 gstabs+, gtoggle, gvms, gxcoff, gxcoff+): Add Driver flag.
1426 2017-01-16 Uros Bizjak <ubizjak@gmail.com>
1428 * config/i386/i386.h (HARD_REGNO_CALLER_SAVE_MODE): Apply HImode and
1429 QImode fixups to general and mask registers only.
1431 2017-01-16 Carl Love <cel@us.ibm.com>
1433 * config/rs6000/rs6000-c (altivec_overloaded_builtins): Add support
1434 for built-in functions
1435 vector signed char vec_nabs (vector signed char)
1436 vector signed short vec_nabs (vector signed short)
1437 vector signed int vec_nabs (vector signed int)
1438 vector signed long long vec_nabs (vector signed long long)
1439 vector float vec_nabs (vector float)
1440 vector double vec_nabs (vector double)
1441 * config/rs6000/rs6000-builtin.def: Add definitions for NABS functions
1443 * config/rs6000/altivec.md: New define_expand nabs<mode>2 types
1444 * config/rs6000/altivec.h: New define for vec_nabs built-in function.
1445 * doc/extend.texi: Update the documentation file for the new built-in
1448 2017-01-16 Martin Sebor <msebor@redhat.com>
1450 * gimple-ssa-sprintf.c (format_directive): Correct a typo in a warning
1453 2017-01-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1455 * config/rs6000/rs6000.c (rtx_is_swappable_p): Change
1456 UNSPEC_VSX__XXSPLTD to require special splat handling.
1458 2017-01-16 David Malcolm <dmalcolm@redhat.com>
1461 * system.h: Poison strndup.
1463 2017-01-16 Alan Modra <amodra@gmail.com>
1466 * config/rs6000/rs6000.c (rs6000_legitimate_combined_insn): Don't
1469 2017-01-16 Georg-Johann Lay <avr@gjlay.de>
1471 * config/avr/avr.h (BRANCH_COST) [reload_completed]: Increase by 4.
1473 2017-01-15 Uros Bizjak <ubizjak@gmail.com>
1475 * config/i386/i386.c (ix86_legitimate_combined_insn): Do not
1476 call recog here. Assert that INSN_CODE (insn) is non-negative.
1478 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1481 * cfgrtl.c (rtl_split_edge): Also patch jump insns that jump to the
1483 * haifa-sched.c (dump_insn_stream): Don't crash if there is a label
1484 in the currently scheduled RTL fragment.
1486 2017-01-15 Segher Boessenkool <segher@kernel.crashing.org>
1488 PR rtl-optimization/78751
1489 * ifcvt.c (find_cond_trap): If we generated a non-existing insn,
1492 2017-01-14 Jeff Law <law@redhat.com>
1494 PR tree-optimization/79090
1495 * tree-ssa-dse.c (valid_ao_ref_for_dse): Reject zero length and
1496 variable length stores.
1497 (compute_trims): Delete dead assignment to *trim_tail.
1498 (dse_dom_walker::dse_optimize_stmt): Optimize mem* calls with
1501 2017-01-14 Bernd Schmidt <bschmidt@redhat.com>
1503 PR rtl-optimization/78626
1504 PR rtl-optimization/78727
1505 * cprop.c (one_cprop_pass): Collect unconditional traps in the middle
1506 of a block, and split such blocks after everything else is finished.
1508 2017-01-14 Alan Modra <amodra@gmail.com>
1511 * combine.c (recog_for_combine_1): Set INSN_CODE before calling
1512 target legitimate_combined_insn.
1513 * config/rs6000/rs6000.c (TARGET_LEGITIMATE_COMBINED_INSN): Define.
1514 (rs6000_legitimate_combined_insn): New function.
1515 * config/rs6000/rs6000.md (UNSPEC_DOLOOP): Delete, and remove
1517 (ctr<mode>_internal3): Rename from *ctr<mode>_internal5.
1518 (ctr<mode>_internal4): Rename from *ctr<mode>_internal6.
1519 (ctr<mode>_internal1, ctr<mode>_internal2): Remove '*' from name.
1521 2017-01-14 Gerald Pfeifer <gerald@pfeifer.com>
1523 * doc/frontends.texi (G++ and GCC): Remove references to Java.
1525 2017-01-13 Jeff Law <law@redhat.com>
1527 PR tree-optimization/33562
1528 PR tree-optimization/61912
1529 PR tree-optimization/77485
1530 * tree-ssa-dse.c (delete_dead_call): Accept gsi rather than
1532 (delete_dead_assignment): Likewise.
1533 (dse_dom_walker::dse_optimize_stmt): Pass in the gsi rather than
1534 statement to delete_dead_call and delete_dead_assignment.
1536 2017-01-13 David Malcolm <dmalcolm@redhat.com>
1539 * substring-locations.c (format_warning_va): Strengthen case 1 so
1540 that both endpoints of the substring must be within the format
1541 range for just the substring to be printed.
1543 2017-01-13 Uros Bizjak <ubizjak@gmail.com>
1545 * config/i386/i386.opt (msgx): Use ix86_isa_flags2 variable.
1546 * config/i386/i386.c (ix86_target_string): Add missing options
1547 to isa_opts and reorder options by implied ISAs. Rename isa_opts2 to
1548 isa2_opts, ix86_flag_opts to flag2_opts, ix86_target_other to
1549 flags_other and ix86_target_other to flags2_other. Display unknown
1551 (ix86_valid_target_attribute_inner_p): Add missing options and
1552 reorder options by implied ISAs, as in ix86_target_string.
1554 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1556 * hash-table.h (hash_table::too_empty_p): New function.
1557 (hash_table::expand): Use it.
1558 (hash_table::traverse): Likewise.
1559 (hash_table::empty_slot): Use sizeof (value_type) instead of
1560 sizeof (PTR) to convert bytes to elements. Shrink the table
1561 if the current size is excessive for the current number of
1564 2017-01-13 Richard Sandiford <richard.sandiford@arm.com>
1566 * ira-costs.c (record_reg_classes): Break from the inner loop
1567 early once alt_fail is known to be true. Update outer loop
1568 handling accordingly.
1570 2017-01-13 Jeff Law <law@redhat.com>
1572 * tree-ssa-dse.c (decrement_count): New function.
1573 (increment_start_addr, maybe_trim_memstar_call): Likewise.
1574 (dse_dom_walker::optimize_stmt): Call maybe_trim_memstar_call directly
1575 when we know the partially dead statement is a mem* function.
1577 PR tree-optimization/61912
1578 PR tree-optimization/77485
1579 * tree-ssa-dse.c: Include expr.h.
1580 (maybe_trim_constructor_store): New function.
1581 (maybe_trim_partially_dead_store): Call maybe_trim_constructor_store.
1583 PR tree-optimization/33562
1584 PR tree-optimization/61912
1585 PR tree-optimization/77485
1586 * doc/invoke.texi: Document new dse-max-object-size param.
1587 * params.def (PARM_DSE_MAX_OBJECT_SIZE): New PARAM.
1588 * tree-ssa-dse.c: Include params.h.
1589 (dse_store_status): New enum.
1590 (initialize_ao_ref_for_dse): New, partially extracted from
1592 (valid_ao_ref_for_dse, normalize_ref): New.
1593 (setup_live_bytes_from_ref, compute_trims): Likewise.
1594 (clear_bytes_written_by, maybe_trim_complex_store): Likewise.
1595 (maybe_trim_partially_dead_store): Likewise.
1596 (maybe_trim_complex_store): Likewise.
1597 (dse_classify_store): Renamed from dse_possibly_dead_store_p.
1598 Track what bytes live from the original store. Return tri-state
1599 for dead, partially dead or live.
1600 (dse_dom_walker): Add constructor, destructor and new private members.
1601 (delete_dead_call, delete_dead_assignment): New extracted from
1603 (dse_optimize_stmt): Make a member of dse_dom_walker.
1604 Use initialize_ao_ref_for_dse.
1606 PR tree-optimization/33562
1607 PR tree-optimization/61912
1608 PR tree-optimization/77485
1609 * sbitmap.h (bitmap_count_bits): Prototype.
1610 (bitmap_clear_range, bitmap_set_range): Likewise.
1611 * sbitmap.c (bitmap_clear_range): New function.
1612 (bitmap_set_range, sbitmap_popcount, bitmap_count_bits): Likewise.
1614 2017-01-13 Martin Liska <mliska@suse.cz>
1617 * function.c (set_cfun): Add new argument force.
1618 * function.h (set_cfun): Likewise.
1619 * ipa-inline-transform.c (inline_call): Use the function when
1620 strict alising from is dropped for function we inline to.
1622 2017-01-13 Richard Biener <rguenther@suse.de>
1624 * tree-pretty-print.c (dump_generic_node): Fix inverted condition
1625 for dumping GIMPLE INTEGER_CSTs.
1627 2017-01-13 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1629 * config/sol2.h (TARGET_OS_CPP_BUILTINS): Define __STDC_VERSION__
1630 to 201112L since C++17.
1632 2017-01-13 Maxim Ostapenko <m.ostapenko@samsung.com>
1635 * asan.c (asan_needs_odr_indicator_p): Don't emit ODR indicators
1636 if -fsanitize=kernel-address is present.
1638 2017-01-13 Richard Biener <rguenther@suse.de>
1640 * tree-pretty-print.c (dump_generic_node): Dump INTEGER_CSTs
1641 as _Literal ( type ) number in case usual suffixes do not
1642 preserve all information.
1644 2017-01-13 Richard Biener <rguenther@suse.de>
1646 PR tree-optimization/77283
1647 * gimple-ssa-split-paths.c: Include gimple-ssa.h, tree-phinodes.h
1648 and ssa-iterators.h.
1649 (is_feasible_trace): Implement a cost model based on joiner
1652 2017-01-12 Michael Meissner <meissner@linux.vnet.ibm.com>
1655 * config/rs6000/rs6000.md (FP_ISA3): Do not optimize converting
1656 char or short to __float128/_Float128 directly.
1658 2017-01-12 Martin Sebor <msebor@redhat.com>
1660 to -Wformat-overflow.
1661 * gimple-ssa-sprintf.c (pass_sprintf_length::gate): Adjust.
1662 (min_bytes_remaining): Same.
1663 (get_string_length): Same.
1664 (format_string): Same.
1665 (format_directive): Same.
1667 (pass_sprintf_length::handle_gimple_call): Same.
1669 2017-01-12 Jakub Jelinek <jakub@redhat.com>
1671 * gimple-ssa-sprintf.c (try_substitute_return_value): Remove
1672 info.nowrite calls with no lhs that can't throw. Return bool
1673 whether gsi_remove has been called or not.
1674 (pass_sprintf_length::handle_gimple_call): Return bool whether
1675 try_substitute_return_value called gsi_remove. Formatting fix.
1676 (pass_sprintf_length::execute): Don't use gsi_remove if
1677 handle_gimple_call returned true.
1680 * cfgrtl.c (rtl_tidy_fallthru_edge): For any_uncondjump_p that can't
1681 be removed due to side-effects, don't remove following barrier nor
1682 turn the successor edge into fallthru edge.
1684 2017-01-12 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1687 * config/rs6000/rs6000.c (insn_is_swappable_p): Mark
1688 element-reversing loads and stores as not swappable.
1690 2017-01-12 Nathan Sidwell <nathan@acm.org>
1691 Nicolai Stange <nicstange@gmail.com>
1693 * combine.c (try_combine): Don't ignore result of overlap checking
1694 loop. Combine overlap & asm check into single loop.
1696 2017-01-12 Richard Biener <rguenther@suse.de>
1698 * tree-pretty-print.c (dump_generic_node): Provide -gimple
1699 variant for MEM_REF. Sanitize INTEGER_CST for -gimple.
1701 2017-01-12 Richard Biener <rguenther@suse.de>
1703 * tree.c (initialize_tree_contains_struct): Make TS_OPTIMIZATION
1704 and TS_TARGET_OPTION directly derive from TS_BASE.
1705 * tree-core.h (tree_optimization_option): Derive from tree_base.
1706 (tree_target_option): Likewise.
1708 2017-01-11 Uros Bizjak <ubizjak@gmail.com>
1710 * config/i386/i386.c (memory_address_length): Increase len
1711 only when rip_relative_addr_p returns false.
1713 2017-01-11 Julia Koval <julia.koval@intel.com>
1715 * common/config/i386/i386-common.c (OPTION_MASK_ISA_SGX_UNSET): New.
1716 (OPTION_MASK_ISA_SGX_SET): New.
1717 (ix86_handle_option): Handle OPT_msgx.
1718 * config.gcc: Added sgxintrin.h.
1719 * config/i386/driver-i386.c (host_detect_local_cpu): Detect sgx.
1720 * config/i386/i386-c.c (ix86_target_macros_internal): Define __SGX__.
1721 * config/i386/i386.c (ix86_target_string): Add -msgx.
1723 (ix86_option_override_internal): Handle new options.
1724 (ix86_valid_target_attribute_inner_p): Add sgx.
1725 * config/i386/i386.h (TARGET_SGX, TARGET_SGX_P): New.
1726 * config/i386/i386.opt: Add msgx.
1727 * config/i386/sgxintrin.h: New file.
1728 * config/i386/x86intrin.h: Add sgxintrin.h.
1730 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1733 * fold-const.c (maybe_nonzero_address): Return 1 for function
1735 (tree_single_nonzero_warnv_p): Don't handle function local objects
1739 * gcc.c (default_compilers): Don't add -o %g.s for -S -save-temps
1742 2017-01-11 David Malcolm <dmalcolm@redhat.com>
1745 * opts.c: Include "spellcheck.h"
1746 (struct string_fragment): New struct.
1747 (struct edit_distance_traits<const string_fragment &>): New
1749 (get_closest_sanitizer_option): New function.
1750 (parse_sanitizer_options): Offer suggestions for unrecognized arguments.
1752 2017-01-11 Jakub Jelinek <jakub@redhat.com>
1754 * dwarf2out.c (DWARF_COMPILE_UNIT_HEADER_SIZE): For DWARF5 decrease
1756 (DWARF_COMDAT_TYPE_UNIT_HEADER_SIZE): Always
1757 DWARF_COMPILE_UNIT_HEADER_SIZE plus 12.
1758 (DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE): Define.
1759 (calc_base_type_die_sizes): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE
1760 for initial die_offset if dwarf_split_debug_info.
1761 (output_comp_unit): Use DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE for
1762 initial next_die_offset if dwo_id is non-NULL. Don't emit padding
1764 (output_skeleton_debug_sections): Formatting fix. Use
1765 DWARF_COMPILE_UNIT_SKELETON_HEADER_SIZE instead of
1766 DWARF_COMPILE_UNIT_HEADER_SIZE. Don't emit padding.
1768 2017-01-11 Wilco Dijkstra <wdijkstr@arm.com>
1770 * config/arm/cortex-a53.md: Add bypasses for
1772 (cortex_a53_r2f): Only use for transfers.
1773 (cortex_a53_f2r): Likewise.
1774 (cortex_a53_r2f_cvt): Add reservation for conversions.
1775 (cortex_a53_f2r_cvt): Likewise.
1777 2017-01-11 Tamar Christina <tamar.christina@arm.com>
1779 * config/arm/arm_neon.h: Add __artificial__ and gnu_inline
1780 to all inlined functions, change static to extern.
1782 2017-01-11 Christophe Lyon <christophe.lyon@linaro.org>
1785 * config/arm/arm.c (legitimize_pic_address): Handle reference to
1787 (arm_assemble_integer): Likewise.
1789 2017-01-11 Richard Earnshaw <rearnsha@arm.com>
1791 * config.gcc: Use new awk script to check CPU, FPU and architecture
1792 parameters for --with-... options.
1793 * config/arm/parsecpu.awk: New file
1794 * config/arm/arm-cpus.in: New file.
1795 * config/arm/arm-opts.h: Include arm-cpu.h instead of processing .def
1797 * config/arm/arm.c: Include arm-cpu-data.h instead of processing .def
1799 * config/arm/t-arm: Update dependency rules.
1800 * common/config/arm/arm-common.c: Include arm-cpu-cdata.h instead
1801 of processing .def files.
1802 * config/arm/genopt.sh: Deleted.
1803 * config/arm/gentune.sh: Deleted.
1804 * config/arm/arm-cores.def: Deleted.
1805 * config/arm/arm-arches.def: Deleted.
1806 * config/arm/arm-fpus.def: Deleted.
1807 * config/arm/arm-tune.md: Regenerated.
1808 * config/arm/arm-tables.opt: Regenerated.
1809 * config/arm/arm-cpu.h: New generated file.
1810 * config/arm/arm-cpu-data.h: New generated file.
1811 * config/arm/arm-cpu-cdata.h: New generated file.
1813 2017-01-11 Maxim Ostapenko <m.ostapenko@samsung.com>
1816 * lto-cgraph.c (lto_output_varpool_node): Pack dynamically_initialized
1818 (input_varpool_node): Unpack dynamically_initialized bit.
1820 2017-01-11 Eric Botcazou <ebotcazou@adacore.com>
1822 PR rtl-optimization/79032
1823 * lra-constraints.c (simplify_operand_subreg): In the MEM case, test
1824 the alignment of the adjusted memory reference against that of MODE,
1825 instead of the alignment of the original memory reference.
1827 2017-01-11 Martin Jambor <mjambor@suse.cz>
1829 * hsa.c (hsa_callable_function_p): Revert addition of DECL_ARTIFICIAL
1831 * ipa-hsa.c (process_hsa_functions): Only duplicate non-artificial
1832 decorated functions.
1834 2017-01-11 Richard Biener <rguenther@suse.de>
1836 * tree-vrp.c (evrp_dom_walker::before_dom_children): Also
1837 set range/nonnull info for PHI results. Do not set it on
1838 stmts marked for removal.
1840 2017-01-10 Eric Botcazou <ebotcazou@adacore.com>
1842 * expr.c (store_field): In the bitfield case, fetch the return value
1843 from the registers before applying a single big-endian adjustment.
1844 Always do a final load for a BLKmode value not larger than a word.
1846 2017-01-10 David Malcolm <dmalcolm@redhat.com>
1849 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
1850 that we correctly handle column numbers greater than
1851 LINE_MAP_MAX_COLUMN_NUMBER.
1853 2017-01-10 Martin Sebor <msebor@redhat.com>
1856 * gimple-ssa-sprintf.c (get_destination_size): Call
1857 {init,fini}object_sizes.
1858 * tree-object-size.c (addr_object_size): Adjust.
1859 (pass_through_call): Adjust.
1860 (pass_object_sizes::execute): Adjust.
1861 * tree-object-size.h (fini_object_sizes): Declare.
1863 2017-01-10 Martin Sebor <msebor@redhat.com>
1865 PR tree-optimization/78775
1866 * builtins.c (get_size_range): Move...
1867 * calls.c: ...to here.
1868 (alloc_max_size): Accept zero argument.
1869 (operand_signed_p): Remove.
1870 (maybe_warn_alloc_args_overflow): Call get_size_range.
1871 * calls.h (get_size_range): Declare.
1873 2017-01-10 Joe Seymour <joe.s@somniumtech.com>
1875 * config/msp430/driver-msp430.c (msp430_mcu_data): Sync with data
1876 from TI's devices.csv file as of September 2016.
1877 * config/msp430/msp430.c (msp430_mcu_data): Likewise.
1879 2017-01-10 Sandra Loosemore <sandra@codesourcery.com>
1881 * doc/extend.texi: Tweak formatting to fix overfull hbox warnings.
1882 * doc/invoke.texi: Likewise.
1883 * doc/md.texi: Likewise.
1884 * doc/objc.texi: Likewise.
1886 2017-01-10 Joshua Conner <joshconner@google.com>
1888 * config/arm/fuchsia-elf.h: New file.
1889 * config/fuchsia.h: New file.
1890 * config.gcc (*-*-fuchsia*): Set native_system_header_dir.
1891 (aarch64*-*-fuchsia*, arm*-*-fuchsia*, x86_64-*-fuchsia*): Add to
1893 * config.host: (aarch64*-*-fuchsia*, arm*-*-fuchsia*): Add to hosts.
1895 2016-01-10 Richard Biener <rguenther@suse.de>
1897 PR tree-optimization/79034
1898 * tree-call-cdce.c (shrink_wrap_one_built_in_call_with_conds):
1899 Propagate out degenerate PHIs in the joiner.
1901 2017-01-10 Martin Liska <mliska@suse.cz>
1903 * ipa-icf.c (sort_sem_items_by_decl_uid): New function.
1904 (sort_congruence_classes_by_decl_uid): Likewise.
1905 (sort_congruence_class_groups_by_decl_uid): Likewise.
1906 (sem_item_optimizer::merge_classes): Sort class, groups in these
1907 classes and members in the groups by DECL_UID of declarations.
1908 This would make merge operations stable.
1910 2017-01-10 Martin Liska <mliska@suse.cz>
1912 * ipa-icf.c (sem_item_optimizer::sem_item_optimizer): Remove
1913 usage of m_classes_vec.
1914 (sem_item_optimizer::~sem_item_optimizer): Likewise.
1915 (sem_item_optimizer::get_group_by_hash): Likewise.
1916 (sem_item_optimizer::subdivide_classes_by_equality): Likewise.
1917 (sem_item_optimizer::subdivide_classes_by_sensitive_refs): Likewise.
1918 (sem_item_optimizer::verify_classes): Likewise.
1919 (sem_item_optimizer::process_cong_reduction): Likewise.
1920 (sem_item_optimizer::dump_cong_classes): Likewise.
1921 (sem_item_optimizer::merge_classes): Likewise.
1922 * ipa-icf.h (congruence_class_hash): Rename from
1923 congruence_class_group_hash. Remove declaration of m_classes_vec.
1925 2017-01-10 Andrew Senkevich <andrew.senkevich@intel.com>
1927 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512VPOPCNTDQ_SET,
1928 OPTION_MASK_ISA_AVX512VPOPCNTDQ_UNSET): New.
1929 * config.gcc: Add avx512vpopcntdqintrin.h.
1930 * config/i386/avx512vpopcntdqintrin.h: New.
1931 * config/i386/cpuid.h (bit_AVX512VPOPCNTDQ): New.
1932 * config/i386/i386-builtin-types.def: Add new types.
1933 * config/i386/i386-builtin.def (__builtin_ia32_vpopcountd_v16si,
1934 __builtin_ia32_vpopcountd_v16si_mask, __builtin_ia32_vpopcountq_v8di,
1935 __builtin_ia32_vpopcountq_v8di_mask): New.
1936 * config/i386/i386-c.c (ix86_target_macros_internal): Define
1937 __AVX512VPOPCNTDQ__.
1938 * config/i386/i386.c (ix86_target_string): Add -mavx512vpopcntdq.
1939 (PTA_AVX512VPOPCNTDQ): Define.
1940 * config/i386/i386.h (TARGET_AVX512VPOPCNTDQ,
1941 TARGET_AVX512VPOPCNTDQ_P): Define.
1942 * config/i386/i386.opt: Add mavx512vpopcntdq.
1943 * config/i386/immintrin.h: Include avx512vpopcntdqintrin.h.
1944 * config/i386/sse.md (define_insn "vpopcount<mode><mask_name>"): New.
1946 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
1949 * predict.def (PRED_CALL): Set to 67.
1951 2017-01-09 Eric Botcazou <ebotcazou@adacore.com>
1953 * expr.c (store_field): In the bitfield case, if the value comes from
1954 a function call and is of an aggregate type returned in registers, do
1955 not modify the field mode; extract the value in all cases if the mode
1956 is BLKmode and the size is not larger than a word.
1958 2017-01-09 Dominique d'Humieres <dominiq@lps.ens.fr>
1961 * config/i386/cpuid.h: Fix undefined behavior.
1963 2017-01-04 Jeff Law <law@redhat.com>
1965 PR tree-optimization/79007
1966 PR tree-optimization/67955
1967 * tree-ssa-alias.c (same_addr_size_stores_p): Only need to be
1968 conservative for pt.null when flag_non_call_exceptions is on.
1970 2017-01-09 Jakub Jelinek <jakub@redhat.com>
1972 PR translation/79019
1973 PR translation/79020
1974 * params.def (PARAM_INLINE_MIN_SPEEDUP,
1975 PARAM_IPA_CP_SINGLE_CALL_PENALTY,
1976 PARAM_USE_AFTER_SCOPE_DIRECT_EMISSION_THRESHOLD): Fix typos
1978 * config/avr/avr.opt (maccumulate-args): Likewise.
1979 * config/msp430/msp430.opt (mwarn-mcu): Likewise.
1980 * common.opt (freport-bug): Likewise.
1981 * cif-code.def (CIF_FINAL_ERROR): Likewise.
1982 * doc/invoke.texi (ipa-cp-single-call-penalty): Likewise.
1983 * config/s390/s390.c (s390_invalid_binary_op): Fix spelling in
1984 translatable string.
1985 * config/i386/i386.c (function_value_32): Likewise.
1986 * config/nios2/nios2.c (nios2_valid_target_attribute_rec): Likewise.
1987 * config/msp430/msp430.c (msp430_option_override, msp430_attr):
1989 * config/msp430/driver-msp430.c (msp430_select_hwmult_lib): Likewise.
1990 * common/config/msp430/msp430-common.c (msp430_handle_option):
1992 * symtab.c (symtab_node::verify_base): Likewise.
1993 * opts.c (set_debug_level): Likewise.
1994 * tree.c (verify_type_variant): Likewise. Fix typo in comment.
1995 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin): Add
1996 missing whitespace to translatable strings.
1997 * config/avr/avr.md (bswapsi2): Fix typo in comment.
1998 * config/sh/superh.h: Likewise.
1999 * config/i386/xopintrin.h: Likewise.
2000 * config/i386/znver1.md: Likewise.
2001 * config/rs6000/rs6000.c (struct rs6000_opt_mask): Likewise.
2002 * ipa-inline-analysis.c (compute_inline_parameters): Likewise.
2003 * double-int.h (struct double_int): Likewise.
2004 * double-int.c (div_and_round_double): Likewise.
2005 * wide-int.cc: Likewise.
2006 * tree-ssa.c (non_rewritable_mem_ref_base): Likewise.
2007 * tree-ssa-sccvn.c (vn_reference_lookup_3): Likewise.
2008 * cfgcleanup.c (crossjumps_occured): Renamed to ...
2009 (crossjumps_occurred): ... this.
2010 (try_crossjump_bb, try_head_merge_bb, try_optimize_cfg, cleanup_cfg):
2013 PR tree-optimization/78899
2014 * tree-if-conv.c (version_loop_for_if_conversion): Instead of
2015 returning bool return struct loop *, NULL for failure and the new
2017 (versionable_outer_loop_p): Don't version outer loop if it has
2018 dont_vectorized bit set.
2019 (tree_if_conversion): When versioning outer loop, ensure
2020 tree_if_conversion is performed also on the inner loop of the
2021 non-vectorizable outer loop copy.
2022 * tree-vectorizer.c (set_uid_loop_bbs): Formatting fix. Fold
2023 LOOP_VECTORIZED in inner loop of the scalar outer loop and
2024 prevent vectorization of it.
2025 (vectorize_loops): For outer + inner LOOP_VECTORIZED, ensure
2026 the outer loop vectorization of the non-scalar version is attempted
2027 before vectorization of the inner loop in scalar version. If
2028 outer LOOP_VECTORIZED guarded loop is not vectorized, prevent
2029 vectorization of its inner loop.
2030 * tree-vect-loop-manip.c (rename_variables_in_bb): If outer_loop
2031 has 2 inner loops, rename also on edges from bb whose single pred
2032 is outer_loop->header. Fix typo in function comment.
2034 2017-01-09 Martin Sebor <msebor@redhat.com>
2037 * asan.c (asan_emit_stack_protection): Increase local buffer size
2038 to avoid snprintf truncation warning.
2040 2017-01-09 Andrew Pinski <apinski@cavium.com>
2042 * config/aarch64/aarch64-cores.def: Add thunderx2t99. Change vulcan
2043 to reference thunderx2t99 for the tuning structure
2044 * config/aarch64/aarch64-cost-tables.h (vulcan_extra_costs):
2046 (thunderx2t99_extra_costs): This.
2047 * config/aarch64/aarch64-tune.md: Regenerate.
2048 * config/aarch64/aarch64.c (vulcan_addrcost_table): Rename to ...
2049 (vulcan_addrcost_table): This.
2050 (vulcan_regmove_cost): Rename to ...
2051 (thunderx2t99_regmove_cost): This.
2052 (vulcan_vector_cost): Rename to ...
2053 (thunderx2t99_vector_cost): this.
2054 (vulcan_branch_cost): Rename to ...
2055 (thunderx2t99_branch_cost): This.
2056 (vulcan_tunings): Rename to ...
2057 (thunderx2t99_tunings): This and s/vulcan/thunderx2t99 .
2058 * doc/invoke.texi (AARCH64/mtune): Add thunderx2t99.
2060 2017-01-09 Martin Jambor <mjambor@suse.cz>
2064 * ipa-prop.h (ipa_jump_func): Swap positions of vr_known and m_vr.
2065 * ipa-cp.c (ipa_vr_operation_and_type_effects): New function.
2066 (propagate_vr_accross_jump_function): Use the above function for all
2067 value range computations for pass-through jump functions and type
2068 converasion from explicit value range values.
2069 (ipcp_propagate_stage): Do not attempt to deduce types of formal
2070 parameters from TYPE_ARG_TYPES.
2071 * ipa-prop.c (ipa_write_jump_function): Remove trailing whitespace.
2072 (ipa_write_node_info): Stream type of the actual argument.
2073 (ipa_read_node_info): Likewise. Also remove trailing whitespace.
2075 2017-01-09 Martin Liska <mliska@suse.cz>
2078 * gcc.c (driver_handle_option): Handle OPT_E and set have_E.
2079 (lookup_compiler): Do not show error message with have_E.
2081 2017-01-09 Jakub Jelinek <jakub@redhat.com>
2083 PR tree-optimization/78938
2084 * tree-vect-stmts.c (vectorizable_condition): For non-masked COND_EXPR
2085 where comp_vectype is VECTOR_BOOLEAN_TYPE_P, use
2086 BIT_{NOT,XOR,AND,IOR}_EXPR on the comparison operands instead of
2087 {EQ,NE,GE,GT,LE,LT}_EXPR directly inside of VEC_COND_EXPR. Formatting
2090 2017-01-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2092 * tree-ssa-address.c (gen_addr_rtx): Don't handle index if it
2095 2017-01-09 Richard Biener <rguenther@suse.de>
2097 PR tree-optimization/78997
2098 * tree-vect-slp.c (vect_mask_constant_operand_p): Handle SSA
2099 name condition properly.
2101 2017-01-09 Richard Biener <rguenther@suse.de>
2104 * dwarf2out.c (is_cxx): New overload with context.
2105 (is_naming_typedef_decl): Use it.
2107 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2109 * invoke.texi (Option Summary): Correct spacing in option lists
2110 and add line breaks to fix over-long lines.
2112 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2116 * extend.texi (Common Variable Attributes): Add xref to GCC
2117 Internals manual to explain mode attribute keywords.
2119 2017-01-08 Sandra Loosemore <sandra@codesourcery.com>
2122 * doc/invoke.texi (Option Summary): Move -pthread to Linker Options
2123 and Preprocessor Options.
2124 (Options for Linking): Document -pthread here....
2125 (RS/6000 and PowerPC Options): ...not here.
2126 (Solaris 2 Options): ...or here.
2127 * doc/cppopts.texi: Document -pthread.
2129 2017-01-08 Martin Sebor <msebor@redhat.com>
2132 * doc/invoke.texi (Warning Options): Document -Wformat-truncation.
2133 * gimple-ssa-sprintf.c (call_info::reval_used, call_info::warnopt):
2134 New member functions.
2135 (format_directive): Used them.
2137 (pass_sprintf_length::handle_gimple_call): Same.
2138 * graphite-sese-to-poly.c (tree_int_to_gmp): Increase buffer size
2139 to avoid truncation for any argument.
2140 (extract_affine_mul): Same.
2141 * tree.c (get_file_function_name): Same.
2143 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2146 * predict.def (PRED_INDIR_CALL): Set to 86.
2148 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2150 PR preprocessor/54124
2151 * doc/cppopts.texi: Reformat -d subtable to list the full name
2152 of the options. Add cross-reference to the docs for the general
2153 compiler -d options.
2154 * doc/invoke.texi (Developer Options): Add cross-reference to the
2155 preprocessor-specific -d option documentation.
2157 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2159 PR preprocessor/13498
2160 * doc/cpp.texi (Search Path): Rewrite to remove obsolete and
2161 redudant material, and reflect new command-line options.
2162 (System Headers): Likewise.
2164 2017-01-07 Sandra Loosemore <sandra@codesourcery.com>
2166 * doc/cppdiropts.texi: Merge documentation of -I, -iquote,
2167 -isystem, and -idirafter. Copy-edit.
2168 * doc/cppopts.texi: Copy-edit. Remove contradiction about
2169 default for -ftrack-macro-expansion. Delete obsolete and
2170 badly-formatted implementation details about -fdebug-cpp output.
2171 * doc/cppwarnopts.texi: Copy-edit.
2173 2017-01-07 David Malcolm <dmalcolm@redhat.com>
2176 * input.c (selftest::test_accessing_ordinary_linemaps): Verify
2177 that the transition from a max line width >= 1<<10 to narrower
2178 lines works correctly.
2180 2017-01-07 Alexandre Oliva <aoliva@redhat.com>
2182 * doc/options.texi (PerFunction): New.
2183 * opt-functions.awk (switch_flags): Map both Optimization and
2184 PerFunction to CL_OPTIMIZATION.
2185 * opth-gen.awk: Test for PerFunction flag along with
2187 * optc-save-gen.awk: Likewise. Introduce var_opt_hash and set
2188 it only when the latter is present. Skip those that don't in
2189 the hash function generator.
2190 * common.opt (fvar-tracking): Mark as PerFunction instead of
2192 (fvar-tracking-assignments): Likewise.
2193 (fvar-tracking-assignments-toggle): Likewise.
2194 (fvar-tracking-uninit): Likewise.
2196 2017-01-07 Jakub Jelinek <jakub@redhat.com>
2198 PR translation/79018
2199 * params.def (PARAM_MAX_STORES_TO_MERGE): Add missing space between
2202 2017-01-06 Mikael Pettersson <mikpelinux@gmail.com>
2205 * config/m68k/m68k.opt (LONG_JUMP_TABLE_OFFSETS): New option.
2206 * config/m68k/linux.h (ASM_RETURN_CASE_JUMP): Handle
2207 TARGET_LONG_JUMP_TABLE_OFFSETS.
2208 * config/m68k/m68kelf.h (ASM_RETURN_CASE_JUMP): Likewise.
2209 * config/m68k/netbsd-elf.h (ASM_RETURN_CASE_JUMP): Likewise.
2210 * config/m68k/m68k.h (CASE_VECTOR_MODE): Likewise.
2211 (ASM_OUTPUT_ADDR_DIFF_ELF): Likewise.
2212 * config/m68k/m68k.md (tablejump expander): Likewise.
2213 (*tablejump_pcrel_hi): Renamed from unnamed insn, reject
2214 TARGET_LONG_JUMP_TABLE_OFFSETS.
2215 (*tablejump_pcrel_si): New insn, handle TARGET_LONG_JUMP_TABLE_OFFSETS.
2216 * doc/invoke.texi (M68K options): Add -mlong-jump-table-offsets.
2218 2017-01-06 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
2219 David Holsgrove <david.holsgrove@xilinx.com>
2221 * common/config/microblaze/microblaze-common.c
2222 (TARGET_EXCEPT_UNWIND_INFO): Remove.
2223 * config/microblaze/microblaze-protos.h (microblaze_eh_return):
2225 * config/microblaze/microblaze.c (microblaze_must_save_register)
2226 (microblaze_expand_epilogue, microblaze_return_addr): Handle
2228 (microblaze_eh_return): New function.
2229 * config/microblaze/microblaze.h (RETURN_ADDR_OFFSET)
2230 (EH_RETURN_DATA_REGNO, MB_EH_STACKADJ_REGNUM)
2231 (EH_RETURN_STACKADJ_RTX, ASM_PREFERRED_EH_DATA_FORMAT): New macros.
2232 * config/microblaze/microblaze.md (eh_return): New pattern.
2234 2017-01-06 Jakub Jelinek <jakub@redhat.com>
2236 * system.h (GCC_DIAGNOSTIC_PUSH_IGNORED, GCC_DIAGNOSTIC_POP,
2237 GCC_DIAGNOSTIC_STRINGIFY): Define.
2239 * read-rtl.c (rtx_reader::read_rtx_code): Avoid -Wsign-compare warning.
2241 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2243 * config/arm/arm.md (<mcrr>): New.
2245 * config/arm/arm.c (arm_arch5te): New.
2246 (arm_option_override): Set arm_arch5te.
2247 (arm_coproc_builtin_available): Add support for mcrr, mcrr2, mrrc
2249 * config/arm/arm-builtins.c (MCRR_QUALIFIERS): Define to...
2250 (arm_mcrr_qualifiers): ... this. New.
2251 (MRRC_QUALIFIERS): Define to...
2252 (arm_mrrc_qualifiers): ... this. New.
2253 * config/arm/arm_acle.h (__arm_mcrr, __arm_mcrr2, __arm_mrrc,
2255 * config/arm/arm_acle_builtins.def (mcrr, mcrr2, mrrc, mrrc2): New.
2256 * config/arm/iterators.md (MCRRI, mcrr, MCRR): New.
2257 (MRRCI, mrrc, MRRC): New.
2258 * config/arm/unspecs.md (VUNSPEC_MCRR, VUNSPEC_MCRR2, VUNSPEC_MRRC,
2259 VUNSPEC_MRRC2): New.
2261 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2263 * config/arm/arm.md (<mcr>): New.
2265 * config/arm/arm.c (arm_coproc_builtin_available): Add
2266 support for mcr, mrc, mcr2 and mrc2.
2267 * config/arm/arm-builtins.c (MCR_QUALIFIERS): Define to...
2268 (arm_mcr_qualifiers): ... this. New.
2269 (MRC_QUALIFIERS): Define to ...
2270 (arm_mrc_qualifiers): ... this. New.
2271 (MCR_QUALIFIERS): Define to ...
2272 (arm_mcr_qualifiers): ... this. New.
2273 * config/arm/arm_acle.h (__arm_mcr, __arm_mrc, __arm_mcr2,
2275 * config/arm/arm_acle_builtins.def (mcr, mcr2, mrc, mrc2): New.
2276 * config/arm/iterators.md (MCRI, mcr, MCR, MRCI, mrc, MRC): New.
2277 * config/arm/unspecs.md (VUNSPEC_MCR, VUNSPEC_MCR2, VUNSPEC_MRC,
2280 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2282 * config/arm/arm.md (*ldc): New.
2286 * config/arm/arm.c (arm_coproc_builtin_available): Add
2287 support for ldc,ldcl,stc,stcl,ldc2,ldc2l,stc2 and stc2l.
2288 (arm_coproc_ldc_stc_legitimate_address): New.
2289 * config/arm/arm-builtins.c (arm_type_qualifiers): Add
2290 'qualifier_const_pointer'.
2291 (LDC_QUALIFIERS): Define to...
2292 (arm_ldc_qualifiers): ... this. New.
2293 (STC_QUALIFIERS): Define to...
2294 (arm_stc_qualifiers): ... this. New.
2295 * config/arm/arm-protos.h
2296 (arm_coproc_ldc_stc_legitimate_address): New.
2297 * config/arm/arm_acle.h (__arm_ldc, __arm_ldcl, __arm_stc,
2298 __arm_stcl, __arm_ldc2, __arm_ldc2l, __arm_stc2, __arm_stc2l): New.
2299 * config/arm/arm_acle_builtins.def (ldc, ldc2, ldcl, ldc2l, stc,
2300 stc2, stcl, stc2l): New.
2301 * config/arm/constraints.md (Uz): New.
2302 * config/arm/iterators.md (LDCI, STCI, ldc, stc, LDC STC): New.
2303 * config/arm/unspecs.md (VUNSPEC_LDC, VUNSPEC_LDC2, VUNSPEC_LDCL,
2304 VUNSPEC_LDC2L, VUNSPEC_STC, VUNSPEC_STC2, VUNSPEC_STCL,
2305 VUNSPEC_STC2L): New.
2307 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2309 * config/arm/arm.md (<cdp>): New.
2310 * config/arm/arm.c (neon_const_bounds): Rename this ...
2311 (arm_const_bounds): ... this.
2312 (arm_coproc_builtin_available): New.
2313 * config/arm/arm-builtins.c (SIMD_MAX_BUILTIN_ARGS): Increase.
2314 (arm_type_qualifiers): Add 'qualifier_unsigned_immediate'.
2315 (CDP_QUALIFIERS): Define to...
2316 (arm_cdp_qualifiers): ... this. New.
2318 (arm_expand_builtin_args): Add case for 6 arguments.
2319 * config/arm/arm-protos.h (neon_const_bounds): Rename this ...
2320 (arm_const_bounds): ... this.
2321 (arm_coproc_builtin_available): New.
2322 * config/arm/arm_acle.h (__arm_cdp): New.
2324 * config/arm/arm_acle_builtins.def (cdp): New.
2326 * config/arm/iterators.md (CDPI,CDP,cdp): New.
2327 * config/arm/neon.md: Rename all 'neon_const_bounds' to
2329 * config/arm/types.md (coproc): New.
2330 * config/arm/unspecs.md (VUNSPEC_CDP, VUNSPEC_CDP2): New.
2331 * gcc/doc/extend.texi (ACLE): Add a mention of Coprocessor intrinsics.
2332 * gcc/doc/sourcebuild.texi (arm_coproc1_ok, arm_coproc2_ok,
2333 arm_coproc3_ok, arm_coproc4_ok): Document new effective targets.
2335 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2337 * config/arm/arm-builtins.c (arm_unsigned_binop_qualifiers): New.
2338 (UBINOP_QUALIFIERS): New.
2340 (acle_builtin_data): New. Change comment.
2341 (arm_builtins): Remove ARM_BUILTIN_CRC32B, ARM_BUILTIN_CRC32H,
2342 ARM_BUILTIN_CRC32W, ARM_BUILTIN_CRC32CB, ARM_BUILTIN_CRC32CH,
2343 ARM_BUILTIN_CRC32CW. Add ARM_BUILTIN_ACLE_BASE and include
2344 arm_acle_builtins.def.
2345 (ARM_BUILTIN_ACLE_PATTERN_START): Define.
2346 (arm_init_acle_builtins): New.
2347 (CRC32_BUILTIN): Remove.
2348 (bdesc_2arg): Remove entries for crc32b, crc32h, crc32w,
2349 crc32cb, crc32ch and crc32cw.
2350 (arm_init_crc32_builtins): Remove.
2351 (arm_init_builtins): Use arm_init_acle_builtins rather
2352 than arm_init_crc32_builtins.
2353 (arm_expand_acle_builtin): New.
2354 (arm_expand_builtin): Use 'arm_expand_acle_builtin'.
2355 * config/arm/arm_acle_builtins.def: New.
2357 2017-01-06 Andre Vieira <andre.simoesdiasvieira@arm.com>
2359 * config/arm/arm-builtins.c (neon_builtin_datum): Rename to ..
2360 (arm_builtin_datum): ... this.
2361 (arm_init_neon_builtin): Rename to ...
2362 (arm_init_builtin): ... this. Add a new parameters PREFIX
2363 and USE_SIG_IN_NAME.
2364 (arm_init_neon_builtins): Replace 'arm_init_neon_builtin' with
2365 'arm_init_builtin'. Replace type 'neon_builtin_datum' with
2366 'arm_builtin_datum'.
2367 (arm_init_vfp_builtins): Likewise.
2368 (builtin_arg): Rename enum's replacing 'NEON_ARG' with
2369 'ARG_BUILTIN' and add a 'ARG_BUILTIN_NEON_MEMORY.
2370 (arm_expand_neon_args): Rename to ...
2371 (arm_expand_builtin_args): ... this. Rename builtin_arg
2372 enum values and differentiate between ARG_BUILTIN_MEMORY
2373 and ARG_BUILTIN_NEON_MEMORY.
2374 (arm_expand_neon_builtin_1): Rename to ...
2375 (arm_expand_builtin_1): ... this. Rename builtin_arg enum
2376 values, arm_expand_builtin_args and add bool parameter NEON.
2377 (arm_expand_neon_builtin): Use arm_expand_builtin_1.
2378 (arm_expand_vfp_builtin): Likewise.
2379 (NEON_MAX_BUILTIN_ARGS): Remove, it was unused.
2381 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2384 * predict.def (PRED_POLYMORPHIC_CALL): Set to 59.
2385 * predict.c (tree_estimate_probability_bb): Reverse direction of
2386 polymorphic call predictor.
2388 2017-01-06 David Malcolm <dmalcolm@redhat.com>
2390 * passes.c (execute_one_pass): Split out pass-skipping logic into...
2391 (determine_pass_name_match): ...this new function and...
2392 (should_skip_pass_p): ...this new function.
2394 2017-01-06 Nathan Sidwell <nathan@acm.org>
2396 * ipa-visibility.c (function_and_variable_visibility): Reformat
2397 comments and long lines. Remove extrneous if.
2398 * symtab.c (symtab_node::make_decl_local): Fix code format.
2399 (symtab_node::set_section_for_node): Fix comment typo.
2401 2017-01-06 Martin Liska <mliska@suse.cz>
2404 * lra-constraints.c: Rename invariant to lra_invariant.
2405 * predict.c (set_even_probabilities): Initialize e to NULL.
2407 2017-01-05 Martin Sebor <msebor@redhat.com>
2409 PR tree-optimization/78910
2410 * gimple-ssa-sprintf.c (tree_digits): Add an argument.
2411 (format_integer): Correct off-by-one error in the handling
2412 of precision with negative numbers in signed conversions..
2414 2017-01-05 Eric Botcazou <ebotcazou@adacore.com>
2416 * doc/invoke.texi (C Dialect Options): Adjust -fsso-struct entry.
2418 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2420 PR tree-optimization/71016
2421 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Pass cond_stmt to
2422 factor_out_conditional_conversion. Formatting fix.
2423 (factor_out_conditional_conversion): Add cond_stmt argument.
2424 If arg1 is INTEGER_CST, punt if new_arg0 is not any operand of
2425 cond_stmt and if arg0_def_stmt is not the only stmt in its bb.
2428 2017-01-05 David Malcolm <dmalcolm@redhat.com>
2430 * Makefile.in (OBJS): Add read-md.o, read-rtl.o,
2431 read-rtl-function.o, and selftest-rtl.o.
2432 * config/aarch64/aarch64.c: Include selftest.h and selftest-rtl.h.
2433 (selftest::aarch64_test_loading_full_dump): New function.
2434 (selftest::aarch64_run_selftests): New function.
2435 (TARGET_RUN_TARGET_SELFTESTS): Wire it up to
2436 selftest::aarch64_run_selftests.
2437 * config/i386/i386.c
2438 (selftest::ix86_test_loading_dump_fragment_1): New function.
2439 (selftest::ix86_test_loading_call_insn): New function.
2440 (selftest::ix86_test_loading_full_dump): New function.
2441 (selftest::ix86_test_loading_unspec): New function.
2442 (selftest::ix86_run_selftests): Call the new functions.
2443 * emit-rtl.c (maybe_set_max_label_num): New function.
2444 * emit-rtl.h (maybe_set_max_label_num): New decl.
2445 * function.c (instantiate_decls): Guard call to
2446 instantiate_decls_1 with if (DECL_INITIAL (fndecl)).
2447 * function-tests.c (selftest::verify_three_block_rtl_cfg): Remove
2449 * gensupport.c (gen_reader::gen_reader): Pass "false"
2450 for new "compact" param of rtx_reader.
2451 * print-rtl.c (rtx_writer::print_rtx_operand): Print "(nil)"
2452 rather than an empty string for NULL strings.
2453 * read-md.c: Potentially include config.h rather than bconfig.h.
2454 Wrap include of errors.h with #ifdef GENERATOR_FILE.
2455 (have_error): New global, copied from errors.c.
2456 (md_reader::read_name): Rename to...
2457 (md_reader::read_name_1): ...this, adding "out_loc" param,
2458 and converting "missing name or number" to returning false, rather
2460 (md_reader::read_name): Reimplement in terms of read_name_1.
2461 (md_reader::read_name_or_nil): New function.
2462 (md_reader::read_string): Handle "(nil)" by returning NULL.
2463 (md_reader::md_reader): Add new param "compact".
2464 (md_reader::read_md_files): Wrap with #ifdef GENERATOR_FILE.
2465 (md_reader::read_file): New method.
2466 * read-md.h (md_reader::md_reader): Add new param "compact".
2467 (md_reader::read_file): New method.
2468 (md_reader::is_compact): New accessor.
2469 (md_reader::read_name): Convert return type from void to file_location.
2470 (md_reader::read_name_or_nil): New decl.
2471 (md_reader::read_name_1): New decl.
2472 (md_reader::m_compact): New field.
2473 (noop_reader::noop_reader): Pass "false" for new "compact" param
2475 (rtx_reader::rtx_reader): Add new "compact" param.
2476 (rtx_reader::read_rtx_operand): Make virtual and convert return
2477 type from void to rtx.
2478 (rtx_reader::read_until): New decl.
2479 (rtx_reader::handle_any_trailing_information): New virtual function.
2480 (rtx_reader::postprocess): New virtual function.
2481 (rtx_reader::finalize_string): New virtual function.
2482 (rtx_reader::m_in_call_function_usage): New field.
2483 (rtx_reader::m_reuse_rtx_by_id): New field.
2484 * read-rtl-function.c: New file.
2485 * selftest-rtl.c (selftest::assert_rtx_ptr_eq_at): New function.
2486 * selftest-rtl.h (ASSERT_RTX_PTR_EQ): New macro.
2487 (selftest::verify_three_block_rtl_cfg): New decl.
2488 * read-rtl-function.h: New file.
2489 * read-rtl.c: Potentially include config.h rather than bconfig.h.
2490 For host, include function.h, memmodel.h, and emit-rtl.h.
2491 (one_time_initialization): New function.
2492 (struct compact_insn_name): New struct.
2493 (compact_insn_names): New array.
2494 (find_code): Handle insn codes in compact dumps.
2495 (apply_subst_iterator): Wrap with #ifdef GENERATOR_FILE.
2496 (bind_subst_iter_and_attr): Likewise.
2497 (add_condition_to_string): Likewise.
2498 (add_condition_to_rtx): Likewise.
2499 (apply_attribute_uses): Likewise.
2500 (add_current_iterators): Likewise.
2501 (apply_iterators): Likewise.
2502 (initialize_iterators): Guard usage of apply_subst_iterator with
2503 #ifdef GENERATOR_FILE.
2504 (read_conditions): Wrap with #ifdef GENERATOR_FILE.
2505 (md_reader::read_mapping): Likewise.
2506 (add_define_attr_for_define_subst): Likewise.
2507 (add_define_subst_attr): Likewise.
2508 (read_subst_mapping): Likewise.
2509 (check_code_iterator): Likewise.
2510 (rtx_reader::read_rtx): Likewise. Move one-time initialization
2512 (one_time_initialization): New function.
2513 (rtx_reader::read_until): New method.
2514 (read_flags): New function.
2515 (parse_reg_note_name): New function.
2516 (rtx_reader::read_rtx_code): Initialize "iterator" to NULL.
2517 Handle reuse_rtx ids.
2518 Wrap iterator lookup within #ifdef GENERATOR_FILE.
2519 Add parsing support for RTL dumps, mirroring the special-cases in
2520 print_rtx, by calling read_flags, reading REG_NOTE names, INSN_UID
2521 values, and calling handle_any_trailing_information.
2522 (rtx_reader::read_rtx_operand): Convert return type from void
2523 to rtx, returning return_rtx. Handle case 'e'. Call
2524 finalize_string on XSTR and XTMPL fields.
2525 (rtx_reader::read_nested_rtx): Handle dumps in which trailing
2526 "(nil)" values were omitted. Call the postprocess vfunc on the
2528 (rtx_reader::rtx_reader): Add new "compact" param and pass to base
2529 class ctor. Initialize m_in_call_function_usage. Call
2530 one_time_initialization.
2531 * rtl-tests.c (selftest::test_uncond_jump): Call
2532 set_new_first_and_last_insn.
2533 * rtl.h (read_rtx): Wrap decl with #ifdef GENERATOR_FILE.
2534 * selftest-rtl.c: New file.
2535 * selftest-rtl.h (class selftest::rtl_dump_test): New class.
2536 (selftest::get_insn_by_uid): New decl.
2537 * selftest-run-tests.c (selftest::run_tests): Call
2538 read_rtl_function_c_tests.
2539 * selftest.h (selftest::read_rtl_function_c_tests): New decl.
2540 * tree-dfa.c (ssa_default_def): Return NULL_TREE for rtl function
2543 2017-01-05 Uros Bizjak <ubizjak@gmail.com>
2545 * config/i386/i386.md (*testqi_ext_3): No need to handle memory
2546 operands in a special way. Assert that pos+len <= mode precision.
2548 2017-01-05 Jakub Jelinek <jakub@redhat.com>
2550 * common.opt (fvect-cost-model): Remove RejectNegative flag, use
2551 3 argument Alias with unlimited for the negative form.
2552 (fno-vect-cost-model): Removed.
2554 2017-01-05 Martin Liska <mliska@suse.cz>
2556 * hsa-gen.c (gen_hsa_divmod): New function.
2557 (gen_hsa_insn_for_internal_fn_call): Use the function for IFN_DIVMOD.
2559 2017-01-05 Martin Liska <mliska@suse.cz>
2562 * gcc.c (lookup_compiler): Reject '-' filename for a precompiled
2565 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2567 * config/s390/s390.c (s390_expand_setmem): Unroll the loop for
2568 small constant length operands.
2570 2017-01-05 Andreas Krebbel <krebbel@linux.vnet.ibm.com>
2572 * config/s390/s390.c (s390_expand_setmem): Avoid overlapping bytes
2573 between loop iterations.
2575 2017-01-05 Martin Liska <mliska@suse.cz>
2578 * gimplify.c (gimplify_decl_expr): Compare to
2579 asan_poisoned_variables instread of checking flags.
2580 (gimplify_target_expr): Likewise.
2581 (gimplify_expr): Likewise.
2582 (gimplify_function_tree): Conditionally initialize
2583 asan_poisoned_variables.
2585 2017-01-04 Jeff Law <law@redhat.com>
2587 PR tree-optimizatin/78812
2588 * rtl.h (contains_mem_rtx_p): Prototype.
2589 * ifcvt.c (containts_mem_rtx_p): Move from here to...
2590 * rtlanal.c (contains_mem_rtx_p): Here and remove static linkage.
2591 * gcse.c (prune_expressions): Use contains_mem_rtx_p to discover
2592 and prune MEMs that are not at the toplevel of a SET_SRC rtx. Look
2593 through ZERO_EXTEND and SIGN_EXTEND when trying to avoid pruning MEMs.
2595 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2597 * input.c (assert_char_at_range): Default-initialize actual_range.
2599 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2601 * df-scan.c (df_ref_create_structure): Make regno unsigned,
2602 to match the caller.
2604 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2606 * cfgexpand.c (expand_gimple_basic_block): Disregard debug
2607 insns after final jump in test to emit dummy move.
2609 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2611 * gimple-iterator.h (gsi_one_nondebug_before_end_p): New.
2612 * tree-eh.c (cleanup_empty_eh): Skip more debug stmts.
2614 2017-01-04 Alexandre Oliva <aoliva@redhat.com>
2616 * multiple_target.c (create_dispatcher_calls): Init e_next.
2617 * tree-ssa-loop-split.c (split_loop): Init border.
2618 * tree-vect-loop.c (vect_determine_vectorization_factor): Init
2621 2017-01-04 Michael Meissner <meissner@linux.vnet.ibm.com>
2626 * config/rs6000/predicates.md (sf_subreg_operand): New predicate.
2627 (altivec_register_operand): Do not return true if the operand
2628 contains a SUBREG mixing SImode and SFmode.
2629 (vsx_register_operand): Likewise.
2630 (vsx_reg_sfsubreg_ok): New predicate.
2631 (vfloat_operand): Do not return true if the operand contains a
2632 SUBREG mixing SImode and SFmode.
2633 (vint_operand): Likewise.
2634 (vlogical_operand): Likewise.
2635 (gpc_reg_operand): Likewise.
2636 (int_reg_operand): Likewise.
2637 * config/rs6000/rs6000-protos.h (valid_sf_si_move): Add declaration.
2638 * config/rs6000/rs6000.c (valid_sf_si_move): New function to
2639 determine if a MOVSI or MOVSF operation contains SUBREGs that mix
2641 (rs6000_emit_move_si_sf_subreg): New helper function.
2642 (rs6000_emit_move): Call rs6000_emit_move_si_sf_subreg to possbily
2643 fixup SUBREGs involving SImode and SFmode.
2644 * config/rs6000/vsx.md (SFBOOL_*): New constants that are operand
2645 numbers for the new peephole2 optimization.
2646 (peephole2 for SFmode unions): New peephole2 to optimize cases in
2647 the GLIBC math library that do AND/IOR/XOR operations on single
2648 precision floating point.
2649 * config/rs6000/rs6000.h (TARGET_NO_SF_SUBREG): New internal
2650 target macros to say whether we need to avoid SUBREGs mixing
2652 (TARGET_ALLOW_SF_SUBREG): Likewise.
2653 * config/rs6000/rs6000.md (UNSPEC_SF_FROM_SI): New unspecs.
2654 (UNSPEC_SI_FROM_SF): Likewise.
2655 (iorxor): Change spacing.
2656 (and_ior_xor): New iterator for AND, IOR, and XOR.
2657 (movsi_from_sf): New insns for SImode/SFmode SUBREG support.
2658 (movdi_from_sf_zero_ext): Likewise.
2659 (mov<mode>_hardfloat, FMOVE32 iterator): Use register_operand
2660 instead of gpc_reg_operand. Add SImode/SFmode SUBREG support.
2661 (movsf_from_si): New insn for SImode/SFmode SUBREG support.
2662 (fma<mode>4): Use gpc_reg_operand instead of register_operand.
2663 (fms<mode>4): Likewise.
2664 (fnma<mode>4): Likewise.
2665 (fnms<mode>4): Likewise.
2666 (nfma<mode>4): Likewise.
2667 (nfms<mode>4): Likewise.
2669 2017-01-04 Marek Polacek <polacek@redhat.com>
2672 * doc/invoke.texi: Document -Wpointer-compare.
2674 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2676 * optc-gen.awk: Emit #error for -W*/-f*/-m* Enum without
2679 * dwarf2out.c (output_loc_list): Don't throw away 64K+ location
2680 descriptions for -gdwarf-5 and emit them as uleb128 instead of
2683 2017-01-04 Kelvin Nilsen <kelvin@gcc.gnu.org>
2686 * doc/sourcebuild.texi (PowerPC-specific attributes): Add
2687 documentation of the powerpc_popcntb_ok attribute.
2688 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
2689 code to issue warning messages if a requested CPU configuration is
2690 not supported by the binary (assembler and loader) toolchain.
2691 (spe_init_builtins): Add two assertions to prevent ICE if attempt is
2692 made to define a built-in function that has been disabled.
2693 (paired_init_builtins): Add assertion to prevent ICE if attempt is
2694 made to define a built-in function that has been disabled.
2695 (altivec_init_builtins): Add comment explaining why definition
2696 of the DST built-in functions is not preceded by an assertion
2697 check. Add assertions to prevent ICE if attempts are made to
2698 define an altivec predicate or an abs* built-in function that has
2700 (htm_init_builtins): Add comment explaining why definition of the
2701 htm built-in functions is not preceded by an assertion check.
2703 2017-01-04 Jeff Law <law@redhat.com>
2705 PR tree-optimizatin/67955
2706 * tree-ssa-alias.c (same_addr_size_stores_p): Check offsets first.
2707 Allow any SSA_VAR_P as the base objects. Use integer_zerop. Verify
2708 the points-to solution does not include pt_null. Use DECL_PT_UID
2711 2017-01-04 Uros Bizjak <ubizjak@gmail.com>
2713 * config/i386/i386.md (HI/SImode test with imm to QImode splitters):
2714 Use gen_int_mode instead of gen_lopwart for const_int operands.
2716 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2718 PR tree-optimization/71563
2719 * match.pd: Simplify X << Y into X if Y is known to be 0 or
2720 out of range value - has low bits known to be zero.
2722 2017-01-04 Alan Modra <amodra@gmail.com>
2724 * Makefile.in (aclocal_deps): Update and order as per aclocal.m4.
2725 * configure: Regenerate.
2726 * config.in: Regenerate.
2728 2017-01-04 Jakub Jelinek <jakub@redhat.com>
2731 * input.c (ebcdic_execution_charset::on_error): Don't use strstr for
2732 a substring of the message, but strcmp with the whole message. Ifdef
2733 ENABLE_NLS, translate the message first using dgettext.
2735 2017-01-03 Jeff Law <law@redhat.com>
2737 PR tree-optimizatin/78856
2738 * tree-ssa-threadupdate.c: Include tree-vectorizer.h.
2739 (mark_threaded_blocks): Remove code to truncate thread paths that
2740 cross multiple loop headers. Instead invalidate the cached loop
2741 iteration information and handle case of a thread path walking
2742 into an irreducible region.
2744 2017-01-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2747 * config/rs6000/rs6000.c (rs6000_split_signbit): Change some
2748 assertions. Add support for doing the signbit if the IEEE 128-bit
2749 floating point value is in a GPR.
2750 * config/rs6000/rs6000.md (Fsignbit): Delete.
2751 (signbit<mode>2_dm): Delete using <Fsignbit> and just use "wa".
2752 Update the length attribute if the value is in a GPR.
2753 (signbit<mode>2_dm_<su>ext): Add combiner pattern to eliminate
2754 the sign or zero extension instruction, since the value is always 0/1.
2755 (signbit<mode>2_dm2): Delete using <Fsignbit>.
2758 * config/rs6000/vsx.md (vsx_extract_<mode>_store_p9): If we are
2759 extracting SImode to a GPR register so that we can generate a
2760 store, limit the vector to be in a traditional Altivec register
2761 for the vextuwrx instruction.
2763 2017-01-03 Ian Lance Taylor <iant@google.com>
2765 * godump.c (go_format_type): Treat ENUMERAL_TYPE like INTEGER_TYPE.
2767 2017-01-03 Martin Sebor <msebor@redhat.com>
2769 PR tree-optimization/78696
2770 * gimple-ssa-sprintf.c (format_floating): Correct handling of
2771 precision. Use MPFR for %f for greater fidelity. Correct handling
2773 (pass_sprintf_length::compute_format_length): Set width and precision
2774 specified by asrerisk to void_node for vararg functions.
2775 (try_substitute_return_value): Adjust dump output.
2777 2017-01-03 David Edelsohn <dje.gcc@gmail.com>
2779 * doc/invoke.texi (RS6000 options): LRA is enabled by default.
2781 2017-01-03 Eric Botcazou <ebotcazou@adacore.com>
2783 * doc/invoke.texi (SPARC options): Document -mlra as the default.
2784 * config/sparc/sparc.c (sparc_option_override): Force LRA unless
2785 -mlra/-mno-lra was passed to the compiler.
2787 2017-01-03 James Cowgill <James.Cowgill@imgtec.com>
2789 PR rtl-optimization/65618
2790 * emit-rtl.c (try_split): Move initialization of "before" and
2791 "after" to just before the call to emit_insn_after_setloc.
2793 2017-01-03 Gerald Pfeifer <gerald@pfeifer.com>
2795 * doc/md.texi (Standard Names): Remove reference to Java frontend.
2797 2017-01-03 Pierre-Marie de Rodat <derodat@adacore.com>
2799 * dwarf2out.c (gen_enumeration_type_die): When
2800 -gno-strict-dwarf, add a DW_AT_encoding attribute.
2802 2017-01-03 Jakub Jelinek <jakub@redhat.com>
2804 PR tree-optimization/78965
2805 * gimple-ssa-sprintf.c (pass_sprintf_length::compute_format_length):
2806 Change first argument from const call_info & to call_info &. For %n
2807 set info.nowrite to false.
2810 * gimple-ssa-sprintf.c (try_substitute_return_value): Don't change
2811 possibly throwing calls.
2813 * genmatch.c (dt_node::gen_kids_1): If generic_exprs include SSA_NAME
2814 and exprs_len || fns_len, emit the code for SSA_NAME next to the exprs
2815 and fns handling, rather than in a separate case SSA_NAME.
2817 2017-01-02 Jeff Law <law@redhat.com>
2819 * config/darwin-driver.c (darwin_driver_init): Const-correctness
2820 fixes for first_period and second_period variables.
2822 2017-01-02 Uros Bizjak <ubizjak@gmail.com>
2825 * config/i386/i386.md (UNSPEC_NOREX_MEM): New unspec.
2826 (*insvqi_1): New insn pattern.
2827 (*insvqi_1_mem_rex64): Ditto.
2829 (*insvqi_3): Rename from *insvqi.
2831 (*extzvqi_mem_rex64): Add UNSPEC_NOREX_MEM tag.
2833 2017-01-02 Gerald Pfeifer <gerald@pfeifer.com>
2835 * doc/cfg.texi (Edges): Remove reference to Java.
2836 (Maintaining the CFG): Ditto.
2838 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2841 * symtab.c (symtab_node::binds_to_current_def_p): Fix handling of
2842 transparent aliases.
2844 2017-01-01 Jan Hubicka <hubicka@ucw.cz>
2847 * predict.def (PRED_CALL): Update hitrate.
2848 (PRED_INDIR_CALL, PRED_POLYMORPHIC_CALL): New predictors.
2849 * predict.c (tree_estimate_probability_bb): Split CALL predictor
2850 into direct/indirect/polymorphic variants.
2852 2017-01-01 Jakub Jelinek <jakub@redhat.com>
2854 Update copyright years.
2856 * gcc.c (process_command): Update copyright notice dates.
2857 * gcov-dump.c (print_version): Ditto.
2858 * gcov.c (print_version): Ditto.
2859 * gcov-tool.c (print_version): Ditto.
2860 * gengtype.c (create_file): Ditto.
2861 * doc/cpp.texi: Bump @copying's copyright year.
2862 * doc/cppinternals.texi: Ditto.
2863 * doc/gcc.texi: Ditto.
2864 * doc/gccint.texi: Ditto.
2865 * doc/gcov.texi: Ditto.
2866 * doc/install.texi: Ditto.
2867 * doc/invoke.texi: Ditto.
2869 Copyright (C) 2017 Free Software Foundation, Inc.
2871 Copying and distribution of this file, with or without modification,
2872 are permitted in any medium without royalty provided the copyright
2873 notice and this notice are preserved.